Sei sulla pagina 1di 40

The Effect of the Kyoto Protocol

on Carbon Emissions

Rahel Aichele
Gabriel Felbermayr

Abstract
Since 1997, CO2 emissions have continued to rise in many countries despite their emission caps under the Kyoto Protocol (Kyoto). Failure to meet promised targets, however,
does not imply that Kyoto has been pointless. Whether Kyoto has made a difference
relative to the counterfactual of No Kyoto is an empirical question that requires an
instrumental variables strategy. We argue that countries ratification of the statutes
governing the International Criminal Court is a valid instrument for ratification of Kyoto commitments. In our panel fixed effects estimations, the instrument easily passes
weak identification and overidentification tests. It can be plausibly excluded from our
second-stage equations and does not cause CO2 emissions. Our estimates suggest that
Kyoto ratification has a quantitatively large (about 10 percent) and robust, though
only moderately statistically significant, negative effect on CO2 emissions. We also
show that higher fuel prices and a different energy mix in Kyoto countries support this
C 2013 by the Association for Public Policy Analysis and Management.
result. 

INTRODUCTION
Arguably, climate change is the most important global environmental problem of
our times. Its policy dimensions are explored by a large and insightful theoretical
literature. Due to the public goods nature of CO2 emissions, it is individually rational
for countries to free ride on others emission reductions. International environmental agreements (IEAs), such as the Kyoto Protocol, exist to solve this dilemma.
In the Kyoto Protocol (Kyoto), 37 industrialized nations and the European Union
(EU) have agreed to cap their levels of greenhouse gas (GHG) emissions to an
average of 94.8 percent of their 1990 emissions by the period 2008 to 2012. Yet since
1997, emissions have continued to rise in many countries despite their emission
caps. In 2010the latest year with GHG data available from the United Nations
Framework Convention on Climate Change (UNFCCC)many countries were still
far from achieving their promised GHG emission reductions. It appears as if the
Kyoto Protocol has been ineffective.
However, our main argument is that failure to meet a promised target does not
imply that Kyoto has been completely unsuccessful in bringing down emissions
relative to the counterfactual situation of No Kyoto (i.e., a counterfactual world
where no Kyoto Protocol exists). One needs to apply program evaluation techniques
to this large-scale policy intervention. Therefore, we ask whether there is empirical
Journal of Policy Analysis and Management, Vol. 32, No. 4, 731757 (2013)

C 2013 by the Association for Public Policy Analysis and Management
Published by Wiley Periodicals, Inc. View this article online at wileyonlinelibrary.com/journal/pam
DOI:10.1002/pam.21720

732 / Kyoto and Carbon Emissions


evidence that the Kyoto Protocol induced emission savings or not. Can an international climate treaty without a strong enforcement mechanism help mitigate climate
change?
To this end, we explore the Kyoto Protocol as an international climate policy
quasi-natural experiment, and ex-post evaluate its effect on emissions. The most
important econometric problem is that selection into Kyoto is most likely not random. The public economics literature argues that GDP per capita, initial emissions,
development status, and political freedom are important determinants of IEA membership, Kyoto including. This complicates correct statistical inference as random
emission shocks are likely to correlate with Kyoto commitments. It is straightforward to control for unobserved time-invariant country-specific correlation by
making use of the panel dimension of the data, but time-specific technology shocks
or changes in environmental preferences could still cause biased estimates. Additionally, emission projections could drive commitment (reverse causation), which
requires an instrumental variables (IV) strategy. Therefore, we explore determinants
of Kyoto membership and contribute an instrument that could potentially be used
in many applications.
In our IV strategy, we use countries membership in the International Criminal
Court (ICC) based in The Hague, Netherlands, as an instrument for Kyoto commitment. Using fixed effects (FEs) estimation, we find robust evidence that Kyoto
commitment reduces CO2 emissions by some 10 percent on average. To corroborate
this surprisingly high effect, we investigate possible channels through which Kyoto
may have affected CO2 emissions. We identify effects on countries energy and electricity mix, fuel prices, and energy and electricity use. We believe that these results
are potentially important for negotiations about future climate deals. They imply
that even a highly imperfect international climate deal may be better than no deal
at all.
The UNFCCC constitutes the legal framework for negotiating international
treaties to reduce GHG emissions. In 1997 such an agreement was signed in the
Japanese city of Kyoto. The Kyoto Protocol entered into force in February 2005,
after Russias ratification ensured the critical 55 percent of world emissions and
55 percent of countries threshold. Aside from a general emission reduction target,
the Kyoto Protocol sets legally binding country-specific targets ranging from an 8
percent reduction for EU countries to a 10 percent increase for Iceland.1 Primarily,
GHG cutbacks have to be achieved with domestic policy measures, but countries
can also use the so-called flexible market mechanisms (emissions trading, clean
development mechanism, and joint implementation) to meet their targets. As in
other IEAs (e.g., the Oslo, Helsinki, or Montreal Protocol), committed countries are
obliged to report various GHGs emissions and their policy measure implementation
status to the UNFCCC. And even though the Kyoto Protocol has an enforcement
body, there is no credible enforcement mechanism.2
A key conceptual question is why a voluntary, nonenforceable agreement such
as Kyoto should solve the prisoners dilemma and matter at all.3 The public economics literature discusses mechanisms through which voluntary IEAs could be
effective. In the case of the Kyoto Protocol, the most compelling argument is that
participating countries have the obligation to monitor and report emissions to the

Taking into account the EU burden sharing agreement, the targets range from a 28 percent reduction
in Luxembourg to a 27 percent increase in Portugal.
2 The WTO is one rare example of an international agreement that provides some enforcement instruments (e.g., in the form of countervailing duties).
3 This question indeed arises for many international institutions and agreements, not just the ones on
global environmental goods.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 733


UNFCCC, which summarizes emission reduction achievements in annual reports.
The resulting league tables are usually the matter of substantial debate and may
well matter for the political process in countries. So, the lack of formal sanctions
has certainly hampered Kyoto, but it does not automatically imply that Kyoto has
not added incentives to engage in mitigation policies with the objective to save
emissions. Whether Kyoto has affected policy and thus emissions is an empirical
question. In this paper, we attempt to provide evidence on this issue.
The rest of the paper proceeds as follows. The next section describes our empirical strategy. First, CO2 emissions are explained by variables capturing economic
development, population growth, political preferences, and trade openness. To account for relevant time-invariant country features such as endowments with natural resources, industrial structure, geographical position, and climate, as well as
for unobserved heterogeneity, we use a FEs estimation strategy. Second, we argue
that countries ICC membership is a valid and relevant instrument. The ICC is a
permanent, international tribunal that prosecutes war crimes and crimes against
humanity. The Rome Statute, which governs the ICC, was adopted in 1998 and
ratified by the necessary quorum of 60 countries at the end of 2002. The Kyoto Protocol was negotiated one year earlier, and has been ratified by countries since 2001.
The timing of the two multilateral initiatives coincided nicely. The two treaties also
posed similar domestic policy issues. For example, commentators such as Groves
(2009) describe both the Kyoto Protocol and the Rome Statute as threats to the
sovereignty of the United States, which has ratified neither. It is thus not surprising that ICC membership correlates robustly to countries commitments under the
Kyoto Protocol. In terms of content, in contrast, the two treaties have nothing in
common. ICC membership does not relate to environmental outcome variables such
as the level of CO2 and GHG emissions or fuel prices, nor is it likely to directly cause
those variables. These features make ICC membership and its spatial lag (i.e., other
countries membership dummies, weighted by their distance and size) candidate
instruments for Kyoto commitment.
The following section presents results from first-stage regressions. We find a robust relationship between ICC and Kyoto membership, both in a linear probability
as well as in a Probit model. The subsequent section presents our core results; it investigates how Kyoto has affected CO2 emissions and presents a host of robustness
checks. Amongst others, we run a placebo check, exclude economies in transition
(EIT) from the sample, use total GHG emissions as an alternative dependent variable, and explore the timing of effects by including lags of Kyoto ratification. The
results show a negative effect of Kyoto commitment on CO2 emissions of around
10 percent. To support this result, the penultimate section turns to the channels
through which Kyoto may have affected emissions. The results suggest that Kyoto
works through changes in the energy mix, higher fuel taxes, and more efficient
energy and electricity use. The last section contains concluding remarks.

RELATED LITERATURE
Our work is related to two important strands of literature. The first group of papers
deals with the economic rationale for IEAs and with their effectiveness. IEAs are
usually understood as ways to solve the free-riding problem inherent in climate
policy. The paper by Congleton (2001) offers a comprehensive survey. Andreoni and
McGuire (1993), Welsch (1995), Hoel (1997), Carraro and Siniscalco (1998), and
Lange and Vogt (2003) describe how international IEAs can solve this dilemma by
providing commitment devices and policy forums for coordination.
However, IEAs (like all international treaties) are based on the voluntary cooperation of countries. This raises the question of whether such agreements can be
Journal of Policy Analysis and Management DOI: 10.1002/pam
Published on behalf of the Association for Public Policy Analysis and Management

734 / Kyoto and Carbon Emissions


effective and how emission targets compliance can be achieved by means of appropriate institutional designs. Gerber and Wichardt (2009) propose a specific example
of an explicit mechanism that induces compliance even in the absence of a central
institution empowered to enforce the agreement. Ringquist and Kostadinova (2005)
and Aakvik and Tjtta (2011) survey the literature and provide more general theoretical arguments on why international agreements can be useful even in the absence
of an enforcement mechanisms. The central reasons are that, first, IEAs could induce scientific research, raise environmental awareness, and change preferences,
thereby affecting technological options and the regulatory environment. Second,
noncompliance with a voluntary IEA could entrain a loss of trust in other international policy arenas like international lending so that countries find it optimal to
comply (Rose & Spiegel, 2009). There is also a growing empirical body of literature
on the efficacy of international institutions. IEAs such as the Montreal and Helsinki
Protocol have been scrutinized, with inconclusive results on their effectiveness to
bring emissions down (see Aakvik & Tjtta, 2011; Finus & Tjtta, 2003; Murdoch,
Sandler, & Vijverberg, 2003; and Ringquist & Kostadinova, 2005).
While focusing on a very specific IEA, the Kyoto Protocol, our work is more
generally related to the literature on international treaties, such as, in the area of
international trade. For example, there is a prominent discussion on whether the
World Trade Organization (WTO) had trade-enhancing effects or not. The initial and
surprisingly inconclusive results by Rose (2004) have triggered important methodological advances; see, for example, Tomz, Goldstein, and Rivers (2007). When
applicable, we make use of the insights of that literature. And the IMF (2009) finds
that voluntary or unenforceable fiscal rules have had significant effects on countries
debt levels. Countries may have dramatically fallen short from proclaimed targets,
but this does not imply that the rules have not had any effect. We contribute to this
strand of literature by investigating the Kyoto Protocol.
A second stream of research deals with the econometric issues that arise due
to the fact that IEA membership or the stringency of climate policy can hardly
be taken as exogenous. Murdoch and Sandler (1997) and Beron, Murdoch, and
Vijverberg (2003) discuss how countries select into such treaties as a function of
their economic development, their demographic situation (see, in particular, York,
2005), or political determinants that simultaneously affect countries emissions and
their willingness to engage in international policy efforts. This is an important
challenge for statistical inference which the literature is only starting to take up.
This challenge looms particularly large when it comes to the quantification of
Kyotos effect on countries CO2 emissions, the most pervasive GHG. Determinants
of emissions are predominantly discussed in the carbon Kuznets curve literature,
which stresses a dynamic relationship between development (measured by GDP
per capita) and CO2 emissions per capita, and started as a purely empirical exercise, see, for example, Grossman and Krueger (1995) and Holtz-Eakin and Selden
(1995) for early contributions. Subsequently, theoretical explanations have been put
forward; Andreoni and Levinson (2001) and Brock and Taylor (2010) provide excellent examples. Dinda (2004) and Galeotti, Lanza, and Pauli (2006) have presented
comprehensive surveys. Much of the Kuznets curve papers attempt estimating the
turning point beyond which further GDP per capita growth lowers emissions per
capita. The evidence for the existence on such a turning point is mixed, and we do
not find conclusive results. Azomahou, Laisney, and Nguyen-Van (2006) present a
recent skeptical view.
More closely related to our work are recent studies that are based on the
Kuznets curve framework, but investigate the role of the Kyoto Protocol. Grunewald
and Martnez-Zarzoso (2009) include a dummy for Kyoto ratification in the carbon Kuznets curve framework. In a panel of 123 countries over the period
1974 to 2004, the authors find that Kyoto commitment reduces CO2 emissions.
Journal of Policy Analysis and Management DOI: 10.1002/pam
Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 735


Iwata and Okada (in press) investigate different types of GHG emissions, and find
Kyoto commitment leads to emission reducing effects for CO2 and CH4 , but not
for NO2 , and positive and significant effects for other GHGs (HFCs, PFCs, SF6 ).
However, both Grunewald and Martnez-Zarzoso (2009) and Iwata and Okada (in
press) do not instrument Kyoto ratification. Aichele and Felbermayr (2012) use an
instrument similar to the one employed in this paper to study the carbon footprints
for a sample of about 40 countries.4
EMPIRICAL STRATEGY AND DATA
A Model of CO2 Emissions: The Second Stage
We are interested in understanding the effect of Kyoto commitments on countries
CO2 emissions. In a later section on transmission channels, we will also investigate
other dependent variables such as the energy mix, fuel prices, and per capita energy
use.
Denoting country is outcome variable of interest at time t by Yit , we want to
estimate the parameter, 1 , in the following relationship
Yit = 0 + 1 Kyotoit + 2 Xit + t + i + it .

(1)

In our core exercise, Yit is the log of CO2 emissions. We also work with a broader
measure of all GHG emissions, but due to data availability, we present results as a
robustness check. Xit is a vector of controls that influence emissions. These controls
can be divided into three different categories. The first set of controls is comprised
of economic and demographic determinants of emissions (the log of GDP and its
square, the log of population, the log of the share of agriculture, manufacturing,
and services in GDP, and economic openness). We expect that an economically
large country has higher levels of emissions, all else equal; population growth increases emissions; and countries with a high share of manufacturing experience
higher levels of emissions, whereas a large share of services and agriculture in GDP
should be associated with fewer emissions. The second category includes measures
for preferences and policy (the stock of other IEAs, a countrys political orientation measured by the chief executives party affiliation, a dummy variable for WTO
membership, and the Polity index).5 The stock of other IEAs is a proxy for environmental awareness. So we expect a country that has signed up for more IEAs to
have lower emissions. Finally, Xit contains the spatial lag of Kyoto commitment,
that is, the Kyoto status of other countries weighted with their respective size over
distance squared. This measure reflects carbon leakage. We expect that countries
with large Kyoto countries nearby (i.e., with a larger value of the spatial Kyoto lag)
are less prone to competitiveness effects and thus have higher own emissions. To
account for time-invariant country-specific determinants such as endowments of
fossil fuels, patterns of comparative advantage, climatic and geographic conditions,

Aichele and Felbermayr (2012) differ from this study in several dimensions: first, the country coverage
is limited to only 40 countries; second, it is more narrowly focused on the proper calculation of carbon
footprints of countries and on their behavior relative to territorial emissions under Kyoto.
5 We do not include an EU dummy. In our fixed effects approach, only the change in EU status matters. In
our time period, with the exception of Malta and Cyprus, new EU members are EIT that have experienced
more changes than just a change in EU status. So, including an EU dummy would not be informative
about the effect of becoming an EU member country. However, in a robustness check we exclude
transition countries from the sample.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

736 / Kyoto and Carbon Emissions


and historical features (such as historical emission levels), we add a full set of country dummies, i . We also include a full set of year dummies, t , to control for the
world business cycle or the world market price of oil.
The key independent variable is a dummy for Kyoto commitment, (K yotoit ). It
takes the value of 1 if a country i has ratified the Protocol at time t and has a cap on
domestic CO2 emissions. It takes the value of 0 otherwise:

Kyotoit =

1 ratification of emission cap and t ratification year,


0
else.

It implies that Kyoto starts to matter for committed countries once ratification
through the parliament has occurred. Studies evaluating the treatment effects of
IEAs such as the Montreal, Helsinki, and Oslo Protocol take a similar stance and
also use ratification as the decisive treatment date (see Ringquist & Kostadinova,
2005).
Econometric Issues
Our default strategy to eliminate the country-specific unobserved heterogeneity, i ,
from equation (1) is FE estimation on yearly data. It is fairly standard and used in
much of the empirical Kuznets curve and IEA treatment effects literature. Yet, this
strategy is subject to some criticism. Bertrand, Duflo, and Mullainathan (2004) argue that standard errors of treatment effects in FE estimation are inconsistent, and
the estimators standard deviation is underestimated if the outcome and treatment
variable are both serially correlated over time. This might lead to an over-rejection
of the Null of no effect. The authors suggest applying a FE estimator to the pretreatment and posttreatment averages (long FEs estimator) to cure the spurious
correlation problem. Although there has been some heterogeneity in the timing of
Kyotos ratification across countries, most countries ratified it between 2001 and
2003. So we assume treatment takes place in this period, but conduct robustness
checks pertaining to this choice. The pretreatment and posttreatment windows of
1997 to 2000 and 2004 to 2007 are chosen such that they constitute symmetric
intervals around this treatment window.6 Whereas the within transformation on
yearly data are subject to the Bertrand, Duflo, and Mullainathans (2004) critique, it
has the advantage that it does not require assumptions about a treatment window.
Moreover, the number of useable observations is about 10 times larger than in the
long FE model. For these reasons we show results for both methods.
FE estimation eliminates any correlation between the second-stage equation errors and the Kyoto dummy that would be due to time-invariant unobserved heterogeneity at the country level. However, such a correlation may arise for two further
reasons. First, if our Kyoto dummy is a noisy measure for a countrys true climate
policy stance, estimates based on a simple within-transformed model would be biased towards zero (attenuation bias). Second, a bias also arises in the presence of
reverse causation, that is, if a transitory shock on the outcome variable makes it
more or less likely that a country has a binding Kyoto commitment. Instrumenting
the Kyoto status cures these biases. Therefore, the key empirical challenge consists
in finding a valid instrument for Kyoto commitment.

Note that Russia and Ukraine ratified Kyoto in 2004 and Belarus in 2005 and are assigned to the
treatment group. Australia and Croatia ratified the Kyoto Protocol in 2007 and are assigned to the
control group.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 737


IV Strategy
In order to identify the causal effect of Kyoto on Yit (emissions and other outcomes),
we require instruments, Zit , that (i) correlate with K yotoit and that (ii) are uncorrelated with the error term in equation (1). Condition (ii) implies that Zit must not
have an effect on Yit except through K yotoit , and that Zit has explanatory power for
K yotoit conditional on the vector of controls, Xit , included into equation (1). An IV
approach then exploits the exogenous variation in instruments for causal inference.
Thus, the first-stage model is


Kyotoit = + Xit + Zit + i + t + it ,

(2)

where i is a country-specific FE and t is a year dummy.


In this paper, we propose ICC membership as an instrument for Kyoto commitment. The Rome Statute, which governs the ICC, was finally signed in 1998, entered
into force in 2002, and by December 2010 ratified by 114 countries. Thirty-four
countries, including the United States, India, and China, decided not to ratify the
Rome Statute. Since the ICC statutes were ratified around the same time as the
Kyoto Protocol, time variation in the ICC membership dummy has the potential to
correlate with time variation in the Kyoto status variable.
The idea behind our instrumentation strategy is that both treaties reflect a countrys preferences for international policy initiatives. Some countries are more willing
than others to give up national sovereignty and subject themselves to an international organization. And indeed, Groves (2009) argues that both the Kyoto Protocol
and the Rome Statute constitute a threat to U.S. sovereignty.7 Further on we test this
linkthat is, condition (i)and estimate the first-stage equation (2) with a linear
probability model.8 If a countrys ICC status is a valid instrument for Kyoto commitment, then its spatial lag (all other countries membership dummies weighted
by a strictly exogenous proxy for trade links) is a valid instrument as well.
The exclusion restriction (ii) cannot be tested formally. It requires that a countrys ICC involvement is not caused by carbon emissions and that it does not directly
affect the outcome variables, either. Meeting these requirements appears plausible enough, but the exclusion restriction also requires that the instrument is not
correlated with the error term. One concern may be that altruistic or cooperative
countries have lower emissions and a higher likelihood of ICC ratification. However, our FE estimation strategy deals with country-specific heterogeneity, which
makes the exclusion restriction more likely to hold. Moreover, we include the stock
of other IEAs that captures how a countrys environmental preferences evolve over
time. We also add variables related to the political orientation of governments, the
polity index, and a WTO dummy to capture any political preferences that may be
related to the ratification of the ICC and Kyoto and also influence emission levels.
Changes in the production structure of a country are taken into account in the
second stage. Components left in the error term of the second-stage equation are
unobservable changes in technological change or changes in comparative advantage
that influence emissions as well as Kyoto commitment. However, this is unrelated to

Similarly, Mike Huckabee, former Governor of Arkansas, argues that the Kyoto Protocol would have
given foreign nations the power to impose standards on us (Huckabee, 2007, p. 70). China expressed
similar concerns in the Copenhagen climate change negotiations. Other developed or newly industrialized
countries that are neither a Kyoto nor an ICC member include Israel, Korea, Singapore, Chile, and Turkey.
8 This appears to be the natural choice since we are working with a linear fixed effects approach in the
second stage. However, the correlation also holds in a Probit framework.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

738 / Kyoto and Carbon Emissions


ICC membership. With these considerations in mind, we argue that we can exclude
ICCit and its spatial lag from the second-stage equation (1).9
Data
We briefly describe the data used in our empirical exercise. CO2 emissions are
from the World Banks World Development Indicators (WDI) 2010. They comprise
emissions due to the burning of fossil fuels and the manufacture of cement, and
include carbon dioxide emissions produced during consumption of solid, liquid,
and gas fuels, and flared gas. Alternatively, we work with a broader measure of GHG
emissions that we also take from WDI. These data are available for most countries
in our sample; however, GHG emissions are not reported on a yearly basis, so we
cannot apply the same econometric strategy when using them.10
Unfortunately, for a sample comprising Kyoto and non-Kyoto countries for a
sufficiently long time window, there is no comprehensive data base on fuel taxes
and, even more desirably, on implicit or explicit carbon prices. However, due to the
nature of our econometric exercise, one can attribute the time variation in fuel prices
to time variation in taxes.11 Data on diesel and gasoline pump prices,12 electricity
and energy use per capita, and the shares of different energy sources in energy and
electricity production were compiled from WDI 2010.
Data on ICC membership stem from the UN Treaty Series database. The ICC
dummy takes a value of 1 if a country has ratified the Rome Statute governing
the ICC and 0 otherwise. The spatial lag of ICC membership is the average ICC
membership of other countries (all other countries membership dummies weighted
by population over distance, and added up). The Kyoto dummy is constructed from
the UNFCCC homepage. Kyotos spatial lag is constructed exactly as the ICC lag.13
GDP, population, and openness data stem from the Penn World Table 6.3. Openness is the usual ratio of exports plus imports over GDP (measured in current prices).
The shares of manufacturing, agriculture, and services (value added) in GDP are obtained from the WDI 2010. The stock of other IEAs was calculated using the IEAs
Database Project.14 It gives the number of IEAs other than Kyoto a country has
ratified or accepted up to a given year. The chief executives party orientation is
from the World Banks Database on Political Institutions (DPI) 2010, and codes the
governments orientation with respect to economic policy as 1 (right wing, conservative, Christian democratic), 2 (centrist), or 3 (left wing, socialist, social democratic,
or communist). A 0 indicates cases where none of the previous categories fit or
the party does not focus on economic issues. The WTO dummy takes a value of
1 if a country is member to the WTO and 0 otherwise, and was compiled from
the WTO homepage. The Polity Index was obtained from the Center of Systemic
Peaces Polity IV Project Database. The index classifies countries according to their

Other international treaties, such as those governing the WTO or international environmental questions, cannot be easily excluded since they will affect emissions directly through green preferences of
voters and consumers, or through trade policy.
10 GHG data are available for the years 1990, 2000, 2005, and 2008. The country coverage ranges from
109 nations in 1990 to 134 for the latest data point.
11 The nonparametric time trend takes care of the global fuel price; country-level fixed effects take care
of endowment-related or geographic determinants of fuel price differences.
12 Pump prices of diesel and gasoline are only available every other year from 1998 to 2006.
13 The exact calculation of the spatial weighting matrix (e.g., with distance squared instead of simple
distance or with all others trade shares, i.e., exports plus imports, in GDP as a more direct measure
of economic competition) does not make a significant difference. Distance data come from the CEPII
distance database.
14 http://iea.uoregon.edu/.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 739


Table 1. Summary statistics.
Variable
Dependent variables, second stage
ln CO2 emissions (kt)
ln GHG emissions (kt of CO2 eq.)
Diesel pump price (USD/L)
Gasoline pump price (USD/L)
Renewables, share in energy use
Alternative energy, share in
electricity production
Fossil fuels, share in energy use
Coal, share in electricity prod.
ln electricity use (kWh/capita)
ln energy use (kg oil eq./capita)
Kyoto variables
Kyoto ratification (0,1)
Kyoto, spatial lag
Kyoto stringency (0,1,2)
Instruments
ICC ratification (0,1)
ICC, spatial lag
Additional controls
ln GDP
ln GDP, squared
ln population
ln manufacturing (in percent of GDP)
ln agriculture (in percent of GDP)
ln service (in percent of GDP)
ln stock of other IEA
Chief executive party orientation
Openness (current price)
WTO dummy (0,1)
Polity index

Obs.

Mean

Std. dev.

Min

Max

Source

1,456
335
622
622
1,209
1,209

9.78
11.20
0.58
0.73
22.64
2.52

2.28
1.56
0.33
0.35
27.45
5.11

4.19
8.39
0.01
0.02
0.00
0.00

15.69
16.04
1.73
1.90
93.77
29.26

(a)
(a)
(a)
(a)
(a)
(a)

1,209
1,209
1,198
1,209

67.82
18.04
7.29
7.17

26.81
27.11
1.49
0.99

5.32 102.43
0.00 99.46
3.02 10.15
4.83
9.38

(a)
(a)
(a)
(a)

1,456
1,456
1,456

0.12
0.28
0.20

0.33
0.97
0.57

0.00
0.00
0.00

1.00
13.29
2.00

(b)
(b)
(b)

1,456
1,456

0.33
0.12

0.47
0.38

0.00
0.00

1.00
3.46

(c)
(c)

1,456 17.95
1,456 325.94
1,456
9.28
1,421
2.62
1,440
2.30
1,437
3.93
1,456
3.28
1,456
0.17
1,456
0.86
1,456
0.82
1,456
3.66

1.91
69.64
1.51
0.56
1.11
0.35
0.59
0.09
0.48
0.39
6.41

13.23 23.28
175.05 542.06
6.04 14.09
0.35
3.79
2.63
4.36
1.01
4.40
1.79
4.84
0.10
0.30
0.05
4.57
0.00
1.00
10.00 10.00

(d)
(d)
(d)
(a)
(a)
(a)
(e)
(f)
(d)
(g)
(h)

Note: The table shows summary statistics for variables over the period 1997 to 2007.
Sources: (a) World Bank WDI 2010, (b) http://www.unfccc.int, (c) UN Treaty Series database, (d) PWT
6.3, (e) http://iea.uoregon.edu, (f) World Bank DPI 2010 (series: execrlc), (g) http://www.wto.org, (h)
http://www.systemicpeace.org.

political authority characteristics and ranges from 10 to 10, where 10 is a perfectly autocratic regime and 10 is a full democracy. Table 1 lists summary statistics
and sources.15
The WDI database has information on emissions for almost 200 geographical
entities. However, we drop entries that do not constitute independent countries
(such as, e.g., Puerto Rico, Greenland, Monaco, and Hong Kong). Moreover, for a
number of countries, we do not have the full list of covariates as defined above. At
this stage we lose mainly small (island) states like Grenada, Barbados, but also some
very recently created countries (such as Kosovo). In our analysis, we drop countries

15

Additional information on Kyoto status, average emission growth from the pretreatment to posttreatment period, and sample info (rich, large, Organization of the Petroleum Exporting Countries (OPEC),
transition country) is given in Appendix Table A1. All appendices are available at the end of this article as
it appears in JPAM online. Go to the publishers Web site and use the search engine to locate the article
at http://www3.interscience.wiley.com/cgi-bin/jhome/34787.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

740 / Kyoto and Carbon Emissions


with fewer than five observations and countries that only appear in either the preor the postpolicy period to get a more balanced sample and have comparability with
the long FE sample. This leaves us with a sample of 133 countries. In 2005, sample
countries account for about 85 percent of global CO2 emissions, which totaled 2.92
billion tons of CO2 . The period of observation covers the years 1997 to 2007. We
chose this time frame because it is symmetric around the treatment time window.
Data availability would allow us to extend the analysis up to 2009; however, the
inclusion of the crisis years 2008 and 2009 does not seem recommendable to us.
SELECTION INTO KYOTO: THE ROLE OF ICC MEMBERSHIP
In our sample, 32 of 133 countries have commitments under the Kyoto Protocol.16
Within the group of countries that had commitments as of 2007, there is some
variation as to the timing of national ratification. The first countries to ratify a
commitment were Romania and the Czech Republic (in 2001), 27 countries ratified
their commitment in 2002, followed by Lithuania (2003), Ukraine (2004), Belarus
(2005), and finally Australia and Croatia (2007). The tetrachoric correlation between
the Kyoto and ICC ratification dummy is 0.78. A two-sided test that Kyoto and ICC
are independent is rejected at all conventional significance levels (P-value 0.00). To
examine what drives Kyoto ratification and verify condition (i), we now estimate
equation (2) with a linear probability model for Kyoto commitments for the time
span 1997 to 2007, applying the same methods (FE and long FE estimation) as for
equation (1).
Table 2 presents results on the first-stage regressions. Columns 1 to 3 report FE
estimations on levels of yearly data. We adjust the variancecovariance matrix for
heteroskedasticity and for clustering of standard errors within countries (Stock &
Watson, 2008). Column 1 shows that ratification of the Rome Statute governing the
ICC correlates strongly with ratification of the Kyoto Protocol. The estimated coefficient of 0.19 implies that ICC ratification increases the odds of Kyoto ratification
by 19 percentage points. The estimate is different from zero at the 1 percent level
of statistical significance and, together with country and time dummies, explains
about 25 percent of the variance in the Kyoto dummy. In column 2, adding the
spatial lag of ICC ratification, that is, ICC ratification by spatially close countries,
further increases the share of variance explained to 52 percent in the FE model.17
Column 3 adds the vector of controls, Xit , which also features in the second-stage
regressions. The spatial lag of Kyoto ratification adds no explanatory power given
the spatial lag of ICC and the other covariates.18 The log of GDP enters negatively,
its square positively (though the latter is not statistically distinguishable from zero).
This signals that economic growth deters countries from ratifying their Kyoto commitment, but the effect eventually levels off. Population size has a large negative
effect on the odds of Kyoto ratification. Since we identify all effects in Table 2 by
time variation at the country level, our results suggest that countries with higher
population growth are less likely to have commitments. This finding is in line with
York (2005), who underlines the importance of demographic factors for the ratification of the Kyoto treaty. Countries industrial structure matters to a certain degree

16

Five committed Kyoto countries (Iceland, Liechtenstein, Luxembourg, Russia, and Switzerland) are
not included due to data availability.
We have also experimented with other international agreements such as the Comprehensive NuclearTest-Ban Treaty and the Anti-Personnel Land Mines Convention. Ratification of these texts also tends to
make Kyoto commitments more likely; however, the effects are weaker and less statistically significant.
18 Note that a positive and statistically significant effect is obtained when the spatial lag of ICC is not
included in the equation.
17

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 741


Table 2. First-stage regressions: explaining Kyoto commitment.
Dependent variable: Kyoto commitment (0,1)
Linear probability
FE

Model
Method

(1)

Excluded instruments
ICC (0,1)
ICC, spatial lag

(2)

0.19*** 0.11***
(0.05)
(0.04)
0.51***
(0.10)

Other controls
Kyoto, spatial lag
ln GDP
ln GDP, squared
ln population
ln manufacturing (percent of GDP)
ln agriculture (percent of GDP)
ln services (percent of GDP)
ln stock of other IEA
Government orientation (0.1, 0.2, 0.3)
Openness, (Exp + Imp)/GDP
WTO (0,1)
Polity (1 to 1)
No. of observations
Adj. R2
F-stat
Log-likelihood
2

1,456
0.25
4.32

1,456
0.52
6.69

Probit long
diff.a
(5)

(3)

Long FE
(4)

0.10***
(0.03)
0.41***
(0.08)

0.25***
(0.07)
0.37***
(0.12)

0.16***
(0.03)
0.04***
(0.01)

0.01
(0.01)
0.94*
(0.52)
0.02
(0.02)
1.72***
(0.38)
0.01
(0.04)
0.14**
(0.06)
0.01
(0.08)
0.02
(0.09)
0.03
(0.19)
0.22***
(0.07)
0.15**
(0.06)
0.00
(0.00)
1,418
0.61
7.36

0.01
(0.01)
2.46**
(1.02)
0.07**
(0.03)
1.31***
(0.45)
0.02
(0.10)
0.29**
(0.13)
0.15
(0.17)
0.27
(0.17)
0.07
(0.59)
0.27
(0.18)
0.28*
(0.14)
0.02
(0.01)
266
0.69
5.69

2.04***
(0.69)
0.05***
(0.02)
3.42***
(0.59)
0.16
(0.12)
0.05
(0.08)
0.25**
(0.13)
0.07
(0.11)
0.11
(0.20)
0.63***
(0.14)
0.16**
(0.07)
0.02***
(0.01)
133
13.92
39.60

Note: Linear probability and Probit models. Sample: 133 countries. Probit regression shows marginal
effects. Heteroskedasticity robust standard errors (clustered at country level) in parentheses. *P < 0.1;
**P < 0.05; ***P < 0.01. LPM: Year dummies and constant included (not shown).
a Estimation on cross-section of differences between pretreatment and posttreatment averages.

for Kyoto commitment: The higher the share of agriculture in a countrys GDP, the
lower the probability it will ratify the Kyoto Protocol. The logs of the manufacturing
and services shares are not relevant for Kyoto commitments.
We include two variables to proxy for green preferences. The first variable, the
log stock of other (than Kyoto) IEAs ratified by a country is expected to affect
the likelihood of commitment positively, but does not show up significantly in
the analysis. The second variable is the countrys chief executive partys political
Journal of Policy Analysis and Management DOI: 10.1002/pam
Published on behalf of the Association for Public Policy Analysis and Management

742 / Kyoto and Carbon Emissions


orientation. One would think that left-leaning governments are more likely to accept
commitments, but this does not show up in our regression.
The next two variables measure openness to international trade: exports plus
imports over GDP and the WTO dummy. Both variables correlate negatively with
Kyoto ratification. On average, doubling openness makes ratification less likely by
22 percentage points. Note that 15 countries became WTO members in the sample
period;19 with the exception of the three Baltic states, none of these countries have
commitments under Kyoto.
As a robustness check, column 4 applies the long FE model, where the change
of Kyoto status (i.e., ratification) is regressed on changes in instruments and other
controls. Comfortingly, the explanatory power of that model is comparable to the
FE model while parameter estimates typically are slightly larger. Column 5 reports
marginal effects pertaining to a Probit model on long-differenced data.20 Results
are again comparable to the long FE linear probability model.
KYOTOS EFFECT ON EMISSIONS
In this section we present our core results about the effect of Kyoto commitment on
CO2 emissions. First, we present the evidence in a graph, and then we employ more
formal econometric FE models. Finally, we turn to several robustness checks.
Graphical Inspection of Kyotos Effects
Figure 1 plots the density of changes in the log of CO2 emissions over two groups of
countries: countries that end up with emission caps, and countries that do not. The
changes are computed over period averages 1997 to 2000 (before the first country
has ratified the Protocol) and 2004 to 2007 (following the ratification of Russia
and Ukraine in 2004). Ratification occurred later only in Belarus (2005), Australia
and Croatia (2007).21 In both groups, emissions have fallen in some countries, but
increased in others. In the sample of non-Kyoto countries, the distribution is more
to the right and the dispersion of changes is much more pronounced. Emissions
have fallen substantially in some developing countries affected by civil war (such as
Burundi), and increased strongly in countries recovering from crises (such as Chad
or Angola). In the group of committed countries, Norway and Spain have increased
emissions by more than 20 percent, while they went down by roughly 8 percent
in Belgium and Germany. Between the two periods, emissions have increased on
average by 27 percent in the group of noncommitted countries, while they have
increased on average by 3.4 percent in the group of committed countries (see the
vertical lines). The difference of the means is statistically significant at the 1 percent
level of significance.
Regression Results
Figure 1 does not control for the effects of time-varying controls and selfselection into the Kyoto Protocol. Therefore, we turn to more elaborate estimation
techniques. Table 3 reports our benchmark results. Column 1 uses ratification of

19

These countries are Albania, Armenia, Cambodia, China, Croatia, Estonia, Georgia, Jordan, Lithuania,
Latvia, Moldova, Macedonia, Nepal, Saudi Arabia, and Vietnam.
20 Time-differencing pre- and postpolicy averages, the selection equation yields a model in which the
dependent variable (Kyoto commitment) and the instrument (ICC) are again binary variables. This
procedure eliminates unobserved heterogeneity; yet, the Probit model remains applicable.
21 Note that Belarus is considered as treated, while Australia and Croatia are put in the control group.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

.01

Density
.02
.03

.04

.05

Kyoto and Carbon Emissions / 743

50

25

25
50
Emission growth (in %)
Kyoto

75

100

NonKyoto

Note: The figure shows a kernel density plot (Epanechnikov, optimal bandwidth) of CO2 emission growth
rates (i.e.,
ln EMit ) of non-Kyoto and Kyoto countries, where t = 0 is the pretreatment (1997 to 2000)
average and t = 1 is the posttreatment (2004 to 2007) average. The graph also shows the mean emission
growth rate in each subpopulation.

Figure 1. Differences in Emission Growth Rates in Kyoto and Non-Kyoto Countries.


Kyoto commitment to explain variation in yearly emissions.22 The Kyoto dummy is
highly statistically significant and suggests that Kyoto commitment reduces emissions by about 17 percent. The simple model explains about 27 percent of the
variation in emission growth across countries.
Column 2 turns to a more comprehensive model. Adding the control vector, Xit ,
more than halves the Kyoto effect and also weakens statistical significance to the
5 percent level.23 The explanatory power of the model rises to about 48 percent.
We find that a 1 standard deviation increase in the spatial lag of Kyoto ratification
(more countries that are close ratify) increases emissions by 1.4 percent.24 This
finding is consistent with the carbon leakage hypothesis: Noncommitted countries
increase emissions as they step up exports of carbon-intensive goods to committed
countries. The coefficients on log GDP and its square are positive and negative, respectively, and not statistically different from zero. Jointly, however, they are highly
significant (the P-value of the F-test statistic of joint significance is 0.00), signaling
a problem of collinearity. Dropping the GDP squared term yields a coefficient on
the log of GDP of 0.56. So, a 1 percent increase of GDP is associated with 0.56 percent higher CO2 emissions. Population increases emissions; the elasticity is highly
significant and statistically not distinguishable from unity.25 This result is in line
with the literature (see, e.g., Cole & Neumayer, 2004). The share of manufacturing in GDP has a positive though not statistically significant effect on emissions,
while a higher share of agriculture or services in GDP reduces CO2 emissions. The
chief executives party (government) orientation has a positive effect on emissions:

22
23

In principle, this replicates Figure 1, but with yearly data.


In the discussion paper version of this paper, we used a larger set of controls and obtained very similar
results.
24 0.965 0.014 = 0.014.
25 The P-value of the F-test statistic is 0.68.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

744 / Kyoto and Carbon Emissions


Table 3. Second-stage regressions: the effect of Kyoto on CO2 emissions.
Dependent variable: ln CO2 emissions

Method
Kyoto (0,1)

FE-OLS
(1)

FE-OLS
(2)

FE-IV
(3)

Long
FE-OLS
(4)

Long
FE-IV
(5)

0.17***
(0.03)

0.06**
(0.02)
0.01**
(0.01)
0.65
(0.55)
0.00
(0.01)
1.10***
(0.24)
0.12
(0.10)
0.09*
(0.05)
0.10
(0.10)
0.18**
(0.07)
0.15**
(0.06)
0.01
(0.07)
0.01
(0.04)
0.00
(0.00)
1,418
133

0.10**
(0.05)
0.02***
(0.00)
0.62
(0.50)
0.00
(0.01)
0.95***
(0.26)
0.12
(0.09)
0.10**
(0.04)
0.10
(0.09)
0.18***
(0.07)
0.15**
(0.06)
0.02
(0.07)
0.00
(0.04)
0.00
(0.00)
1,418
133

0.09**
(0.04)
0.03***
(0.01)
1.13**
(0.54)
0.02
(0.01)
1.12***
(0.26)
0.25***
(0.09)
0.26***
(0.08)
0.26*
(0.15)
0.23**
(0.10)
0.29**
(0.15)
0.16
(0.12)
0.04
(0.05)
0.01
(0.01)
266
133

0.12*
(0.07)
0.03***
(0.01)
1.02*
(0.56)
0.01
(0.02)
1.05***
(0.29)
0.25***
(0.09)
0.27***
(0.09)
0.26*
(0.15)
0.21*
(0.11)
0.28*
(0.15)
0.16
(0.12)
0.03
(0.06)
0.01
(0.01)
266
133

Kyoto, spatial lag


ln GDP
ln GDP, squared
ln population
ln manufacturing (percent of GDP)
ln agriculture (percent of GDP)
ln services (percent of GDP)
ln stock of other IEA
Government orientation (0.1,0.2,0.3)
Openness, (Exp + Imp)/GDP
WTO (0,1)
Polity (1 to 1)
No. of observations
No. of countries
First-stage diagnostics
Sheas partial R2
HansenSargan J-stat (P-value)
Weak-ID test (F-stat)
Second-stage diagnostics
Adj. R2
F-stat

1,456
133

0.28
0.44
19.09
0.27
11.99

0.48
13.85

17.44

0.43
0.71
37.70
0.49
21.40

23.09

Note: Standard errors in parentheses adjusted for within-group clustering and heteroskedasticity. *P <
0.1; **P < 0.05; ***P < 0.01. Instruments for Kyoto variable: ICC membership and its spatial lag. All
regressions use a comprehensive set of year dummies (not reported).

Left-leaning governments are associated with higher emissions. Openness appears


to lower emissions, but the effect is not statistically significant. The WTO dummy is
also insignificant. Political freedom (the polity variable) does not affect emissions
either.
Column 3 applies our IV strategy, using both the ICC dummy and its spatial lag as
instruments. This strategy turns out successful: The weak identification test yields a
KleibergenPaapWald F-statistic of 19.09. This is comfortingly above the canonical
10 percent level suggested by Staiger and Stock (1997) and gives a maximal IV size
bias of 15 percent as tabulated by Stock and Yogo (2005). This is supported by
Sheas partial R2 of 0.28. About 28 percent of the variation in the Kyoto dummy
Journal of Policy Analysis and Management DOI: 10.1002/pam
Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 745


is explained by the instruments only. The overidentification test yields a Hansen
J-statistic of 0.59 with the associated 2 (1) P-value of 0.44. Hence, the Null that the
instruments are uncorrelated with the error term is not rejected and the instruments
appear valid. Results on the controls differ only slightly across the ordinary least
squares (OLS) and IV models.26 Instrumentation slightly increases the estimated
Kyoto effect; this feature is present in most of our IV regressions. The estimated
coefficient implies that Kyoto commitment caused a reduction of CO2 emissions of
10 percent compared to the counterfactual of No Kyoto.
To meet the Bertrand, Duflo, and Mullainathan (2004) critique of spurious correlation due to serial correlation between treatment and outcome variable, we also
use the long FE estimator. The long FE estimates on average pretreatment and
posttreatment data yield very similar results to the yearly FE model. The sign and
significance of covariates is mostly unaffected. In the long FE model, the positive effect of GDP growth on emissions is now significant, and also the industrial structure
matters for emissions. First-stage diagnostics signal instrument validity. The Kyoto
coefficients obtained in both the OLS model in column 5 and the instrumented
model in column 6 are slightly larger, but estimated at a somewhat smaller degree
of precision. Yet, the finding of a negative effect of Kyoto commitment on emissions
does not reflect a purely spurious correlation.
Robustness Checks
Table 4 summarizes robustness checks on our IV regressions. We provide results
of a placebo test, present estimates based on alternative IV estimators, modify the
definition of the Kyoto treatment, work with different subsamples, propose an alternative treatment window, investigate a different outcome variable, and incorporate
time lags of Kyoto ratification. We focus on the Kyoto coefficient; full regression
results are in Appendix Table A2.27
Column A applies a placebo treatment to the group of Kyoto countries in the
prepolicy period (i.e., 1997 to 2000); assuming, counterfactually, that ratification
occurred in 1999 instead of between 2002 and 2004, it reports results from a FE
model. The placebo treatment does not turn out to be statistically significant. We
conclude that our baseline results are not driven by trends in CO2 in the control and
treatment groups that have been present in the data before Kyoto ratification took
place.
Columns B1 and B2 show that the benchmark results of Table 3 are robust to using
different IV strategies. Using limited information maximum likelihood (LIML) or a
two-step generalized method of moments (GMM) estimation yields identical results
to employing two-stage least squares.28
Column C suggests that the baseline results are robust to defining Kyoto commitment in a somewhat finer fashion: The Kyoto stringency variable takes the value
of 0 if a country has no obligations under Kyoto, a value of 1 if the country has a

26

We do not report adjusted R2 for IV regressions because they have no statistical meaning (e.g.,
Wooldridge, 2012, p. 523).
27 All appendices are available at the end of this article as it appears in JPAM online. Go to the
publishers Web site and use the search engine to locate the article at http://www3.interscience.
wiley.com/cgi-bin/jhome/34787.
28 LIML is a linear combination of the FE and the FE-IV estimator (with the weights depending on the
data), and the weights happen to be such that they (approximately) eliminate the FE-IV bias resulting
from weak instruments. In theory, LIML has better small sample properties, so that differences in
estimates would reflect the presence of a bias.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

746 / Kyoto and Carbon Emissions


Table 4. Robustness checks: summary table IV estimates on emissions.
(A) Placebo
treatment
1999
(A)
Kyoto (0,1)
No. of observations
No. of countries
Over-ID test (P-value)
Weak-ID test (F-stat)

0.06
(0.04)
518
133

(B) Alternative
IV strategy
LIML
(B1)

GMM
(B2)

(C) Kyoto
definition
Stringency
(C)

0.10**
(0.05)
1,418
133
0.44
19.09

0.10**
(0.04)
1,418
133
0.44
19.09

0.05**
(0.02)
1,418
133
0.41
21.94

(D) Alternative samples


Sample
Kyoto (0,1)
No. of observations
No. of countries
Over-ID test (P-value)
Weak-ID test (F-stat)

W/o EIT
(D1)

W/o OPEC
(D2)

Rich only
(D3)

Large only
(D4)

0.10*
(0.05)
1,161
109
0.36
17.08

0.10**
(0.05)
1,354
127
0.67
18.18

0.11***
(0.04)
720
67
0.10
8.30

0.09
(0.06)
1,007
94
0.07
16.28

(E) Alternative
treatment window

Kyoto (0,1)
No. of observations
No. of countries
Over-ID test (P-value)
Weak-ID test (F-stat)

(F) GHGs

Narrow
(E1)

Broad
(E2)

2000,
2005
(F1)

2000,
2008
(F2)

0.12**
(0.06)
266
133
0.85
34.06

0.17**
(0.07)
258
129
0.82
37.47

0.10***
(0.03)
218
109
0.48
10.33

0.27***
(0.10)
200
100
0.41
11.59

Note: Dependent variable is ln CO2 emissions, in Panel F, ln GHG emission. FE estimation. In Panels
E and F, long FE estimation. Standard errors in parentheses adjusted for within-group clustering and
heteroskedasticity. *P < 0.1; **P < 0.05; ***P < 0.01. Panel A imposes placebo treatment in 1999 for the
prepolicy period 1997 to 2000. Panels B to F: instruments for Kyoto: ICC membership and its spatial lag.
All regressions use year dummies and additional controls: ln GDP; ln GDP, squared; ln population; ln
IEAs; Chief executive party orientation; openness; WTO dummy; polity. Full results in Appendix Table
A2.29

cap that does not bind (as of 2007), and a value of 2 if it has a binding cap. The
IV strategy continues to work nicely; the Kyoto effect is statistically significant and
quantitatively comparable to the baseline estimations.30

29

All appendices are available at the end of this article as it appears in JPAM online. Go to the publishers Web site and use the search engine to locate the article at http://www3.interscience.wiley.com/
cgi-bin/jhome/34787.
30 Note that the support of the stringency variable is in [0,2] instead of in [0,1].

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 747


Column D1 excludes EIT in Europe and the former Soviet Union from the sample.31 The IV strategy remains valid. The effect of Kyoto remains negative and
of similar size to earlier regressions. The effect is statistically significant at the
10 percent level. Hence, we find that excluding transition countries does not undo
the negative effect of Kyoto on emissions. Results are broadly robust to excluding
Organization of the Petroleum Exporting Countries (OPEC) countries in column D2
or focusing on the 50 percent richest countries (in terms of per capita income as
of 2007) in column D3.32 Limiting attention to countries with more than 5 million
inhabitants in column D4 leaves the magnitude of the Kyoto effect unchanged, but
it is not statistically significant. And in this case, the IV strategy is not successful;
instrument endogeneity is rejected at the 7 percent level.33
Panel E turns once more to long FE models and asks whether our definition of the
treatment window influences the results. Almost 90 percent of the countries ratified
in 2002. So, column E1 uses only 2002 as the treatment year.34 Column E2 defines
a broad treatment window from 2000 to 2004.35 The results are robust to these
modifications and broadly in line with the earlier long FE results both in terms of
magnitude and significance.
Columns F1 and F2 use a comprehensive measure of GHG emissions instead of
CO2 emissions. Since GHG data are only available in certain years, we use a long
FE approach. Comparing the year 2000 (before ratification) to the years 2005 or
2008 (after ratification), we confirm a strongly significant negative effect of Kyoto
on emissions.36
The effect of Kyoto on emissions could evolve over time and be cumulative due
to regulatory lags. To take this possibility into account, we incorporate up to three
time lags of the Kyoto ratification dummy in Model 1.37 Results are reported in Table 5, column 1. Jointly, Kyoto and its lags are significant (the P-value of the F-test
statistic is 0.04). All Kyoto coefficients are negative, but only the immediate effect
and the third lag are statistically significant at conventional levels. The significant
effects add up to the effect obtained in our benchmark. We indeed find that changes
in CO2 emissions following Kyoto ratification take time because policies have to be
implemented and processes have to be adapted. This is also supported by an alternative estimation where we include a dummy for Kyotos entry into force alongside
the ratification dummy (not shown). Both ratification and entry into force enter
negatively, with 0.03 (significant at the 10 percent level) and 0.06 (significant at
the 1 percent level) as point estimates, respectively. The main policy action seems
to start with entry into force. However, we also observe some anticipation effects.

31

The definition of EIT follows the IMF (2000). The excluded countries are Albania, Armenia, Azerbaijan,
Belarus, Bulgaria, Croatia, Czech Republic, Estonia, FYR of Macedonia, Georgia, Kazakhstan, Kyrgyzstan, Latvia, Lithuania, Poland, Republic of Moldova, Romania, Slovak Republic, Slovenia, Tajikistan,
Turkmenistan, Ukraine, and Uzbekistan. We do not treat Germany as a transition country.
32 The F-stat of 8.3 in column (D3) signals a possible weak instrument problem. However, LIML estimation, as suggested by Stock and Yogo (2005) for such cases, shows that the point estimate of 0.11 is
robust and statistically significant at the 5 percent level. The maximal LIML size bias is 15 percent.
33 This finding may reflect treatment heterogeneity: If small countries are more likely to accept ambitious
binding commitments or are more willing to effectively implement them, their presence in the sample is
necessary for a negative average treatment effect.
34 Kyoto countries that ratified after 2002 are still considered to be treated. The only exceptions being
Australia and Croatia, which ratified Kyoto in late 2007.
35 Only Belarus ratification lies outside the broad treatment window.
36 Note, however, that the sample shrinks to 109 and 100 countries due to data availability.
37 Due to the short time horizon of our analysis, there is a natural limit to the number of lags one can
meaningfully include. We experimented with several combinations. The fourth lag is not statistically
significant and lowered the F-statistic of a test of joint significance of Kyoto and its time lags. Also note
that it is not possible to instrument these lags, so estimates are to be interpreted with care.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

748 / Kyoto and Carbon Emissions


Table 5. Kyoto effects with lag structure.
Share of

Dep. var.
Kyoto (0,1)
Kyoto1 (0,1)
Kyoto2 (0,1)
Kyoto3 (0,1)
No. of
observations
No. of countries
Adj. R2
F-stat
F-test joint
significance
(P-value)

(1)
CO2
emission

(2)
Renewables

(3)
Fossils

0.03*
(0.02)
0.00
(0.01)
0.02
(0.01)
0.04***
(0.02)

0.80**
(0.39)
0.42**
(0.21)
0.26
(0.20)
0.57
(0.36)

1,418
133
0.48
13.28
0.038

Price of

ln of per capita use of

(4)
Coal

(5)
New
Clean

(6)
Diesel

(7)
Gasoline

(8)
Energy

(9)
Electricity

0.56
(0.53)
0.42
(0.39)
0.28
(0.46)
0.94**
(0.45)

0.03
(0.67)
0.80
(0.72)
0.28
(0.60)
0.93*
(0.53)

0.50***
(0.18)
0.07
(0.24)
0.57**
(0.24)
1.23***
(0.31)

0.01
(0.03)
0.05
(0.10)
0.14
(0.10)
0.03
(0.03)

0.05
(0.03)
0.05
(0.07)
0.14**
(0.07)
0.05
(0.03)

0.03*
(0.01)
0.00
(0.01)
0.02**
(0.01)
0.05***
(0.01)

0.03
(0.02)
0.01
(0.01)
0.01
(0.01)
0.03*
(0.02)

1,180

1,180

1,180

1,180

610

610

1,180

1,169

110
0.29
6.11
0.173

110
0.17
5.74
0.127

110
0.21
2.73
0.000

129
0.79
57.50
0.000

129
0.70
51.12
0.000

110
0.39
7.48
0.000

109
0.52
14.25
0.405

110
0.06
1.86
0.067

Note: FE estimations with time lags. Standard errors in parentheses adjusted for within-group clustering and heteroskedasticity. *P < 0.1; **P < 0.05; ***P < 0.01. All regressions use year dummies and additional controls: ln GDP; ln
GDP, squared; ln population; ln share of manufacturing/services/agriculture; ln IEAs; Chief executive party orientation;
openness; WTO dummy; polity (not reported). Full results in Appendix Table A3.38

Summarizing, we employ our instruments in a second-stage regression that estimates the effect of Kyoto commitment on CO2 emissions. Our instruments perform
very well: The weak identification test yields an F-statistic of at least 16.28 and often
substantially higher; the overidentification tests are easily passed. The effect of Kyoto ratification on emissions is robustly negative over all econometric models and
survives a host of robustness checks. The effect is economically substantial: Kyoto
has caused CO2 emissions to fall relative to the counterfactual by about 10 percent.
However, the estimates are not always very precise (though typically different from
zero at the 10 percent to 5 percent level). We find a similar effect for GHG emissions.
CHANNELS: FUEL MIX, FUEL PRICES, AND ELECTRICITY USE
Is a negative relationship between Kyoto commitment and CO2 emissions plausible?
The Kyoto Protocol impacts emissions through domestic policies and measures
such as cap-and-trade systems for GHGs (in the EU, Norway, Switzerland, and
New Zealand), changes in fuel taxes, feed-in tariffs for alternative energy sources,
regulations (of, e.g., cars fuel economy, insulation requirements for houses, low
energy bulbs), and energy efficiency and mandatory performance standards for
electrical products. In the following subsections, we present evidence that sheds
light on the channels through which Kyoto may have led to lower emissions. We
focus on energy and electricity mix, fuel prices, as well as energy and electricity use.
The selection of channels is certainly exploratory, and it is mostly driven by data
availability.39

38

All appendices are available at the end of this article as it appears in JPAM online. Go to the publishers Web site and use the search engine to locate the article at http://www3.interscience.wiley.com/
cgi-bin/jhome/34787.
39 For example, there is not sufficient data to investigate the channels of regulation or labeling.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 749

.2

1.5

New clean (##)

.15

.2

.1

10

10

.5

10 5

10

15

20

10

10

20

10

Diesel price

Gasoline price

Ln energy/capita

Ln electricity/capita

40

60

US cents

80

.03
.01

50

50

100

US cents

20

.02

.02

.01

.01

.02

.02

.04

.03

.04

.05

percentage points

.06

percentage points

.04

percentage points

percentage points

.03

20

.05
0

.05

.1

.1

.2

.15

.3

Coal (##)

Fossil fuels (#)


.25

.4

Renewables (#)

40

20

Kyoto

20

40

50

50

100 150 200

NonKyoto

Note: The diagrams show kernel density plots (Epanechnikov, optimal bandwidth) of changes in respective variables (i.e.,
Yit ) of non-Kyoto and Kyoto countries, where t = 0 is the pretreatment (1997 to
2000) average and t = 1 is the posttreatment (2004 to 2007) average. New clean refers to wind and solar
energy. #: Share in total energy use. ##: Share in electricity production.

Figure 2. Changes in Energy Mix, Fuel Prices, Energy Use, and Kyoto Commitment.

However, changes in the energy mix could indicate that Kyoto works through
feed-in tariffs or other ways to promote green energy (e.g., R&D subsidies for green
energy sources). Likewise, changes in fuel prices will pick up effects from fuel
taxes. On the other hand, changes in energy and electricity use are not directly
related to one specific policy measure, but rather reflect changes along all policy
dimensions combined. But these outcome variables help to disentangle technique
and composition from scale effects (see, e.g., Copeland & Taylor, 2003). Hence,
we can learn whether emission mitigation results from a switch toward less CO2 intensive production technologies (technique effect) and changes in the composition
of produced goods (composition effect)both resulting in a lower energy intensity
or whether adjustments occur along the scale of production (scale effect).
Evidence for effects along these channels would support our findings in the section
on Kyotos effect on emissions. As discussed for the case of CO2 emissions, the same
endogeneity issues arise. For example, countries with a higher share of wind, solar,
or nuclear energy may find it easier to commit to Kyoto. Therefore, we provide FE
OLS and IV results based on the same models and instrumentation strategy.
Energy Mix
We start with a graphical inspection of the effect of Kyoto commitments on countries energy mix. The upper leftmost diagram in Figure 2 shows a kernel density plot
Journal of Policy Analysis and Management DOI: 10.1002/pam
Published on behalf of the Association for Public Policy Analysis and Management

750 / Kyoto and Carbon Emissions


of changes in the share of renewables and waste in total energy use40 between the
pretreatment and posttreatment period. On average, committed countries increased
this share by about 1 percentage point, while it decreased by about 2.4 percentage
points in the sample of noncommitted countries. The difference of 3.58 percentage points is statistically significant at the 1 percent level. The next diagram shows
changes in the share of fossil fuel41 in total energy use. The difference between the
group averages is 2.85 percentage points, statistically significant at the 1 percent
level. It appears, thus, that changes in the fossil fuel share correlate negatively with
Kyoto commitment.
Table 6 provides FE regression results on yearly data.42 Columns A1 and A2 investigate the renewables share. The OLS effect implies that Kyoto commitment
increases that share by 1.38 percentage points. Instrumentation almost doubles the
effect of Kyoto to 2.41, statistically significant at the 1 percent level. ICC membership and its spatial lag continue to be good instruments: Both the overidentification
and the weak instrument tests suggest instrument validity. Columns A3 to A4 examine the fossil fuel share in energy use. The OLS estimate of Kyoto is 0.67 and not
significant. Overall, it is not very easy to explain changes in the share of fossil fuel,
probably due to a lack of time variance in the dependent and the independent variables. The IV point estimate is negative and statistically significant at the 5 percent
level. It implies that Kyoto commitment results in a 2.46 percentage point reduction
in the fossil fuel share. So compared to the counterfactual, Kyoto countries rely less
on fossil fuel, and thus CO2 -intensive energy sources. The upshot of these findings
is that, at least partly, the emissions savings brought about by Kyoto commitment
have been achieved by countries switching their energy mix away from fossil fuels
toward renewables.
Electricity Mix
Next, we turn to time changes (again, between 1997 and 2007) in the share of
coal and new clean forms of energy (i.e., wind and solar) in electricity production. The two right-hand diagrams in the upper row of Figure 2 provide illustrations. Changes in the share of coal use are distributed around a mean of close
to zero in both the sample of Kyoto countries and in the sample of uncommitted countries, suggesting that Kyoto commitment probably has not had any measurable impact on coal use. Interestingly, the distribution of changes is much
more dispersed for non-Kyoto countries. The upper rightmost diagram explores
the alternative energy share. Interestingly, there are a few noncommitted countries where wind and solar energy expanded substantially (Kenya, El Salvador,
Nicaragua), but in the sample of noncommitted countries as a whole, the share
only increased by 0.5 percentage points. In the sample of committed countries,
Spain, Portugal, and Germany have considerably increased their shares of new
clean energy sources. The group average is 2 percentage points. The growth difference between the two groups is 1.68 percentage points, significant at the 1 percent
level.

40
41
42

This energy source comprises solid and liquid biomass, biogas, industrial, and municipal waste.
Fossil fuel comprises coal, oil, petroleum, and natural gas products.
To save space, Table 6 only reports Kyoto estimates and first- and second-stage diagnostics. Full
results are delegated to Appendix Table A4. All appendices are available at the end of this article as it
appears in JPAM online. Go to the publishers Web site and use the search engine to locate the article at
http://www3.interscience.wiley.com/cgi-bin/jhome/34787.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 751


Table 6. Through which channels does Kyoto operate? Summary table OLS and IV estimates.
Panel

(A) Shares in energy use

Dep. var.

Renewables

Fossil fuel

(B) Shares in electricity


production
Coal

Alternative
energy

(A1)
(A2)
(A3)
(A4)
(B1)
(B2)
(B3)
(B4)
FE-OLS FE-IV FE-OLS FE-IV FE-OLS FE-IV FE-OLS FE-IV

Method
Kyoto (0,1)
No. of observations
No. of countries
Sheas partial R2
Over-ID test (P-value)
Weak-ID test (F-stat)
Adj. R2
Panel

1.38**
(0.56)
1,180
110

2.41***
(0.93)
1,180
110
0.26
0.68
18.80

0.29

0.67
(0.63)
1,180
110

2.46**
(1.16)
1,180
110
0.26
0.47
18.80

0.17

Diesel fuel

1.43
(1.76)
1,180
110
0.26
0.60
18.80

0.06

(C) Pump prices (USD/L)

Dep. var.

0.12
(0.93)
1,180
110

Gasoline

1.07***
(0.27)
1,180
110

1.66***
(0.56)
1,180
110
0.26
0.63
18.80

0.17

(D) Log per capita use of


Energy

Electricity

(C1)
(C2)
(C3)
(C4)
(D1)
(D2)
(D3)
(D4)
FE-OLS FE-IV FE-OLS FE-IV FE-OLS FE-IV FE-OLS FE-IV

Method
Kyoto (0,1)
No. of observations
No. of countries
Sheas partial R2
Over-ID test (P-value)
Weak-ID test (F-stat)
Adj. R2

0.10**
(0.04)
608
127

0.77

0.22***
(0.05)
608
127
0.28
0.61
20.30

0.13***
(0.04)
608
127

0.68

0.25***
(0.05)
608
127
0.28
0.32
20.30

0.05** 0.05*
(0.02) (0.03)
1,180
1,180
110
110
0.26
0.36
18.80
0.38

0.04
(0.03)
1,169
109

0.08**
(0.04)
1,169
109
0.26
0.21
18.84

0.52

Note: Standard errors in parentheses adjusted for within-group clustering and heteroskedasticity.
*P < 0.1; **P < 0.05; ***P < 0.01. Instruments for Kyoto: ICC membership and its spatial lag. All regressions use year dummies and additional controls: ln GDP; ln GDP, squared; ln population; ln share of
manufacturing/services/agriculture; ln IEAs; Chief executive party orientation; openness; WTO dummy;
polity (not reported). Full results in Appendix Table A4.43

Panel B of Table 6 shows regression results. Columns B1 and B2 analyze the coal
share in electricity production. Controlling for a host of variables and applying our
IV strategy, there is no statistical evidence in favor of a negative effect of Kyoto on
the coal share. This is as suggested by our graphical analysis. One may view this as a
surprising result since coal is the most CO2 intensive fuel for electricity production.
However, it seems that countries have preferred to cut back on the use of other
fossil fuels, especially petrol, which is more likely to be imported, therefore meeting
less local political resistance. Columns B3 and B4 study the share of alternative
energy sources. The IV regression suggests that Kyoto commitments have increased
that share by about 1.66 percentage points. It appears that CO2 -free energy sources

43

All appendices are available at the end of this article as it appears in JPAM online. Go to the publishers Web site and use the search engine to locate the article at http://www3.interscience.wiley.com/
cgi-bin/jhome/34787.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

752 / Kyoto and Carbon Emissions


have gained ground in Kyoto countries electricity generation, thus lowering CO2
emissions all else equal.

Fuel Prices
We continue with the fuel price channel. The two lower left panels of Figure 2 plot
kernel densities of the (absolute) changes in fuel prices, expressed in U.S. cents per
liter, across the groups of committed and noncommitted countries. Between the
pretreatment and posttreatment period, the diesel price has increased by about 30
cents in the group of noncommitted countries and by 49 in committed countries.
The differential increase across the two groups is 18 cents. It is significant at the
1 percent level. A similar picture emerges when looking at gasoline. The average
increase in noncommitted countries was 29 cents and in committed countries 45.
The difference, 16 cents, is again statistically significant at the 1 percent level. Note
that comparing changes over time between Kyoto and non-Kyoto countries, we are
effectively controlling for changes in the world market prices for these fuels.
Panel C in Table 6 confirms these patterns. Column C1 reports the uninstrumented
FE estimator. It suggests that Kyoto status is positively associated to the price of
diesel fuel. With an adjusted R2 of 77 percent, the specification is surprisingly successful in predicting the diesel price. Instrumenting Kyoto commitment leaves the
controls virtually unchanged, but the Kyoto effect more than doubles to 22 cents per
liter. The IV strategy turns out to work reasonably well: The weak identification test
yields an F-statistic of 20.3, and the overidentification test does not reject instrument
validity. The results for the price of gasoline (columns C3 and C4) look very similar.
Again, the IV estimation yields a Kyoto point estimate that is about double the OLS
estimate. In summary, Kyoto commitment has increased fuel prices considerably.
Given the design of our econometric exercise where year dummies control for the
world price of the fuel and country FE control for endowments and proximity to
these endowments, we may interpret these findings as indirect evidence that Kyoto
commitment has led to higher fuel taxes.

Energy and Electricity Use


Finally, we are interested in the role of Kyoto commitment for the per capita use
of energy and electricity. Again, we show the distribution of changes for Kyoto and
non-Kyoto countries; see the two lower panels on the right in Figure 2. Perhaps not
surprisingly, the two distributions are substantially more concentrated for Kyoto
countries. Also mean changes appear to differ across the two groups of countries.
Energy use per capita (in kg of oil equivalent) has increased by about 4 percentage
points less in the group of committed countries. The effect is statistically significant
at the 5 percent level. Turning to the log electricity per capita consumption (in kWh)
in the rightmost diagram, cross-country variation in growth rates is wider than for
energy use per capita. Electricity use per capita has grown in almost all countries.
The average growth rate over the two observed periods is 28 percent in the sample of
noncommitted countries and 12 percent in the group of committed countries. The
difference, 16 percentage points, is statistically significant at the 1 percent level.
Panel D in Table 6 reports regression results. Column D1 suggests that Kyoto commitment reduces per capita energy use by 5 percent. When instrumenting for Kyoto,
the negative relationship is unchanged, but estimated with less statistical precision.
The evidence is different for electricity consumption per capita. Kyoto commitment has no statistically significant effect without instrumentation, but turns out to
reduce energy use per capita by about 8 percent in the IV model.
Journal of Policy Analysis and Management DOI: 10.1002/pam
Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 753


Robustness Checks
Table 7 summarizes robustness checks pertaining to the IV estimates of Table 6.
We apply long FE estimation and vary the sample by excluding transition countries
or focusing on rich countries only. In general, these modifications do not qualitatively affect the benchmark results. However, the evidence on the change of energy
and electricity use due to Kyoto commitment is not very robust. The IV strategy
continues to work with three exceptions.44 And for all regressions based on rich
countries only, the F-statistic of the weak identification test is very low. Using LIML
estimation remedies this problem and reduces the maximum LIML size bias to 15
percent for all specifications, which is not overly large.
Table 5 documents that Kyotos effect on the energy mix, fuel prices, and energy
and electricity use accumulates over time. Typically, the sign structure is robust
over the lags. The sum of significant coefficients adds up to roughly the long-run
effects estimated in our baseline regression. Also, the overall fit of the regressions is
not substantially improved by including lags, see the R2 statistics.
Summarizing, using the same IV strategy as for CO2 emissions, we find that Kyoto has increased the share of renewables in energy production by about 2 to 3
percentage points and reduced that of fossil fuels by similar amounts. It has also
increased the share of alternative energy in electricity production by about 2 percentage points. Relative to the counterfactual, prices of diesel and gasoline are about
20 U.S. cents higher in Kyoto countries. Finally, Kyoto appears to have lowered per
capita electricity and energy use, although this finding is less robust. These results
are robust to excluding transition countries. These channels contribute to emission
savings and thus support the finding of Kyoto countries emission reductions.
DISCUSSION
We provide robust evidence that Kyoto commitment reduced CO2 emissions. We
also identify channels through which Kyoto may have affected emissions. Our paper fits in the larger literature on the empirical evaluation of the effectiveness of
IEAs. As pointed out, previous studies on the Helsinki or Oslo Protocol regulating SO2 emissions do not find effects. These studies acknowledge the self-selection
into treaties problem and use Heckman-type self-selection models (Ringquist &
Kostadinova, 2005) or random-effect and FE estimation (Aakvik & Tjtta, 2011;
Ringquist & Kostadinova, 2005). However, they are conceptually different from our
approach in that they do not employ IV estimation and are, therefore, prone to
attenuation and reverse causality biases. Furthermore, the Kyoto Protocol might be
special in at least two respects. First, the issue of global warming is discussed very
controversially and trails a huge public debate. So, informal enforcement mechanisms could play a more important role. Second, CO2 emissions are associated
with costly inputs: energy and electricity. There is always an economic incentive to
save on those. Our paper certainly cannot reconcile opposing views on voluntary
IEAs effectiveness. Future research is needed to assess the general effectiveness of
IEAs.
In our study, we only shortly touch the issue of carbon leakage. We saw that lower
fuel prices and a change in energy and electricity generation may explain part of the
observed emission reductions. Yet, our study is silent on another possible channel:
relocation of production to non-Kyoto countries and reexport to Kyoto countries.

44

The overidentification test signals endogenous instruments in the long FE estimation for diesel prices,
the FE model on rich countries only for gasoline prices, and the specification without EITs for per capita
electricity use.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

754 / Kyoto and Carbon Emissions


Table 7. Robustness checks: channelssummary table IV estimates.
Method/sample

Long FE

W/O EIT

Rich only

Long FE

W/O EIT

Rich only

(A) Shares in energy use


Renewables
Dep. var.
Kyoto (0,1)
No. of observations
Over-ID test (P-value)
Weak-ID test (F-stat)

Fossil fuel

(A1)

(A2)

(A3)

(A4)

(A5)

(A6)

3.43***
(1.14)
220
0.32
36.49

3.44***
(1.29)
923
0.53
15.81

2.05**
(0.85)
701
0.72
8.39

3.62***
(1.37)
220
0.18
36.49

2.80**
(1.29)
923
0.50
15.81

0.58
(1.13)
701
0.58
8.39

(B) Shares in electricity production


Alternative energy
Dep. var.
Kyoto (0,1)
No. of observations
Over-ID test (P-value)
Weak-ID test (F-stat)

Coal

(B1)

(B2)

(B3)

(B4)

(B5)

(B6)

3.06***
(0.58)
220
0.60
36.49

2.06***
(0.56)
923
0.55
15.81

1.69***
(0.54)
701
0.30
8.39

0.96
(1.41)
220
0.62
36.49

1.71
(2.29)
923
0.47
15.81

0.74
(1.30)
701
0.92
8.39

(C) Pump prices


Diesel fuel
Dep. var.
Kyoto (0,1)
No. of observations
Over-ID test (P-value)
Weak-ID test (F-stat)

Gasoline

(C1)

(C2)

(C3)

(C4)

(C5)

(C6)

0.49***
(0.06)
252
0.04
39.06

0.12**
(0.05)
498
0.67
18.86

0.31***
(0.08)
316
0.39
8.83

0.46***
(0.06)
252
0.47
39.06

0.17***
(0.06)
498
0.34
18.86

0.32***
(0.09)
316
0.04
8.83

(D) Per capita use of


Energy
Dep. var.
Kyoto (0,1)
No. of observations
Over-ID test (P-value)
Weak-ID test (F-stat)

Electricity

(D1)

(D2)

(D3)

(D4)

(D5)

(D6)

0.06
(0.04)
220
0.67
36.49

0.07*
(0.04)
923
0.31
15.81

0.06*
(0.03)
701
0.42
8.39

0.02
(0.06)
218
0.92
36.30

0.13***
(0.04)
912
0.03
16.23

0.06
(0.04)
701
0.74
8.39

Note: Standard errors in parentheses adjusted for within-group clustering and heteroskedasticity. *P <
0.1; **P < 0.05; ***P < 0.01. Instruments for Kyoto: ICC membership and its spatial lag. All regressions
use year dummies and additional controls as in Table 6. Full results in Appendix Table A5.45

45

All appendices are available at the end of this article as it appears in JPAM online. Go to the
publishers Web site and use the search engine to locate the article at http://www3.interscience.
wiley.com/cgi-bin/jhome/34787.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 755


This would also generate the observed pattern of fewer emissions in Kyoto countries.
Assessing the empirical importance of carbon leakage is another important avenue
for future research.
RAHEL AICHELE is a postdoctoral researcher at the IFO InstituteLeibniz Institute for Economic Research at the University of Munich, Poschingerstrae 5, 81679
Munich, Germany.
GABRIEL FELBERMAYR is a Professor of Economics at Ludwig-MaximiliansUniversitat
Munich, Geschwister-Scholl-Platz 1, 80539 Munich, Germany.
ACKNOWLEDGMENTS
We are grateful to Peter Egger, Mario Larch, Mary Lovely, Devashish Mitra, David Popp, M.
Scott Taylor, Martin Wagner, and to participants at the NoEG, EEA and VfS annual meetings
in 2011 for comments, and to the German Science Foundation (DFG) for financial support
(grant no. 583467).

REFERENCES
Aakvik, A., & Tjtta, S. (2011). Do collective actions clear common air? The effect of international environmental protocols on sulphur emissions. European Journal of Political
Economy, 27, 343351.
Aichele, R., & Felbermayr, G. (2012). Kyoto and the carbon footprint of nations. Journal of
Environmental Economics and Management, 63, 336354.
Andreoni, J., & Levinson, A. (2001). The simple analytics of the environmental Kuznets curve.
Journal of Public Economics, 80, 269286.
Andreoni, J., & McGuire, M. C. (1993). Identifying the free riders: A simple algorithm for
determining who will contribute to a public good. Journal of Public Economics, 51, 447
454.
Azomahou, T., Laisney, F., & Nguyen-Van, P. (2006). Economic development and CO2 emissions: A nonparametric panel approach. Journal of Public Economics, 90, 13471363.
Beron, K. J., Murdoch, J. C., & Vijverberg, W. P. M. (2003). Why cooperate? Public goods,
economic power, and the Montreal protocol. Review of Economics and Statistics, 85, 286
297.
Bertrand, M., Duflo, E., & Mullainathan, S. (2004). How much should we trust differencesin-differences estimates? Quarterly Journal of Economics, 119, 249275.
Brock, W. A., & Taylor, M. S. (2010). The green Solow model. Journal of Economic Growth,
15, 127153.
Carraro, C., & Siniscalco, D. (1998). International institutions and environmental policy: International environmental agreements: Incentives and political economy. European Economic Review, 42, 561572.
Cole, M. A., & Neumayer, E. (2004). Examining the impact of demographic factors on air
pollution. Population & Environment, 26, 521.
Congleton, R. (2001). Governing the global environmental commons: The political economy of
international environmental treaties and institutions. In G. Schulze & H. Ursprung (Eds.),
Globalization and the environment (pp. 241263). New York, NY: Oxford University Press.
Copeland, B. R., & Taylor, M. S. (2003). Trade and the environment: Theory and evidence.
Princeton and Oxford: Princeton University Press.
Dinda, S. (2004). Environmental Kuznets curve hypothesis: A survey. Ecological Economics,
49, 431455.
Finus, M., & Tjtta, S. (2003). The Oslo protocol on sulfur reduction: The great leap forward?
Journal of Public Economics, 87, 20312048.
Journal of Policy Analysis and Management DOI: 10.1002/pam
Published on behalf of the Association for Public Policy Analysis and Management

756 / Kyoto and Carbon Emissions


Galeotti, M., Lanza, A., & Pauli, F. (2006). Reassessing the environmental Kuznets curve for
CO2 emissions: A robustness exercise. Ecological Economics, 57, 152163.
Gerber, A., & Wichardt, P. C. (2009). Providing public goods in the absence of strong institutions. Journal of Public Economics, 93, 429439.
Grossman, G. M., & Krueger, A. B. (1995). Economic growth and the environment. Quarterly
Journal of Economics, 110, 353377.
Groves, S. (2009). The Kyoto II climate change treaty: Implications for American
sovereignty. Heritage Foundation Special Report No. 72. Washington, DC: The Heritage
Foundation. Retrieved June 3, 2013, from http://www.heritage.org/research/reports/2009/
11/the-kyoto-ii-climate-change-treaty-implications-for-american-sovereignty.
Grunewald, N., & Martnez-Zarzoso, I. (2009). Driving factors of carbon dioxide emissions
and the impact from Kyoto protocol. CESifo Working Paper No. 2758. Munich, Germany: CESifo Group. Retrieved June 3, 2013, from http://www.cesifo-group.de/pls/
guestci/download/CESifo%20Working%20Papers%202009/CESifo%20Working%
20Papers%20August%202009/cesifo1_wp2758.pdf.
Hoel, M. (1997). Coordination of environmental policy for transboundary environmental
problems? Journal of Public Economics, 66, 199224.
Holtz-Eakin, D., & Selden, T. M. (1995). Stoking the fires? CO2 emissions and economic
growth. Journal of Public Economics, 57, 85101.
Huckabee, M. (2007). From hope to higher ground: 12 steps to restoring Americas greatness.
New York, NY: Center Street.
IMF (2000). Transition economies: An IMF perspective on progress and prospects. Washington, DC: International Monetary Fund. Retrieved June 3, 2013, from http://www.
imf.org/external/np/exr/ib/2000/110300.htm.
IMF (2009). Fiscal rulesAnchoring expectations for sustainable public finances. IMF staff
paper. Washington, DC: International Monetary Fund. Retrieved June 3, 2013, from http://
www.imf.org/external/np/pp/eng/2009/121609.pdf.
Iwata, H., & Okada, K. (in press). Greenhouse gas emissions and the role of the Kyoto
protocol. Environmental Economics and Policy Studies.
Lange, A., & Vogt, C. (2003). Cooperation in international environmental negotiations due to
a preference for equity. Journal of Public Economics, 87, 20492067.
Murdoch, J. C., & Sandler, T. (1997). The voluntary provision of a pure public good: The case
of reduced CFC emissions and the Montreal protocol. Journal of Public Economics, 63,
331349.
Murdoch, J. C., Sandler, T., & Vijverberg, W. P. (2003). The participation decision versus the
level of participation in an environmental treaty: A spatial Probit analysis. Journal of Public
Economics, 87, 337362.
Ringquist, E. J., & Kostadinova, T. (2005). Assessing the effectiveness of international environmental agreements: The case of the 1985 Helsinki protocol. American Journal of Political
Science, 49, 86102.
Rose, A. K. (2004). Do we really know that the WTO increases trade? American Economic
Review, 94, 98114.
Rose, A. K., & Spiegel, M. M. (2009). Noneconomic engagement and international exchange:
The case of environmental treaties. Journal of Money, Credit and Banking, 41, 337363.
Staiger, D., & Stock, J. H. (1997). Instrumental variables regression with weak instruments.
Econometrica, 65, 557586.
Stock, J. H., & Watson, M. W. (2008). Heteroskedasticity-robust standard errors for fixed
effects panel data regression. Econometrica, 76, 155174.
Stock, J. H., & Yogo, M. (2005). Testing for weak instruments in linear IV regression. In D.
W. K. Andrews & J. H. Stock (Eds.), Identification and inference for econometric models
Essays in honor of Thomas Rothenberg (pp. 80108). Cambridge, MA: Cambridge University Press.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions / 757


Tomz, M., Goldstein, J. L., & Rivers, D. (2007). Do we really know that the WTO increases
trade? American Economic Review, 97, 20052018.
Welsch, H. (1995). Incentives for forty-five countries to join various forms of carbon reduction
agreements. Resource and Energy Economics, 17, 213237.
Wooldridge, J. M. (2012). Introductory econometrics: A modern approach (5th ed.). Mason,
OH: South-Western College Publisher.
York, R. (2005). Kyoto protocol participation: A demographic explanation. Population Research and Policy Review, 24, 513526.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions


APPENDIX
Table A1. Country and sample information.

ISO3

Country

AGO
ALB
ARE
ARG
ARM
AUS
AUT
AZE
BDI
BEL
BEN
BFA
BGD
BGR
BLR
BOL
BRA
BTN
BWA
CAN
CHL
CHN
CIV
CMR
COG
COL
COM
CRI
CUB
CYP
CZE
DEU
DJI
DNK
DOM
DZA
EGY
ERI
ESP
EST
ETH
FIN
FJI
FRA
GAB
GIN
GMB
GNQ

Angola
Albania
United Arab Emirates
Argentina
Armenia
Australia
Austria
Azerbaijan
Burundi
Belgium
Benin
Burkina Faso
Bangladesh
Bulgaria
Belarus
Bolivia
Brazil
Bhutan
Botswana
Canada
Chile
China
Cote dIvoire
Cameroon
Congo, Rep.
Colombia
Comoros
Costa Rica
Cuba
Cyprus
Czech Republic
Germany
Djibouti
Denmark
Dominican Republic
Algeria
Egypt, Arab Rep.
Eritrea
Spain
Estonia
Ethiopia
Finland
Fiji
France
Gabon
Guinea
The Gambia
Equatorial Guinea

Sample
Emission growth Kyoto ICC
(in percent)
(year) (year) Rich Large OPEC EIT
91
66
30
18
28
8
13
11
55
8
81
36
44
2
10
19
9
32
23
11
12
55
3
33
27
3
40
32
3
16
1
6
12
5
6
22
30
12
23
9
16
11
72
1
10
7
34
261

n.a.
n.a.
n.a.
n.a.
n.a.
2007
2002
n.a.
n.a.
2002
n.a.
n.a.
n.a.
2002
2005
n.a.
n.a.
n.a.
n.a.
2002
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
2001
2002
n.a.
2002
n.a.
n.a.
n.a.
n.a.
2002
2002
n.a.
2002
n.a.
2002
n.a.
n.a.
n.a.
n.a.

n.a.
2003
n.a.
2001
n.a.
2002
2000
n.a.
2004
2000
2002
2004
n.a.
2002
n.a.
2002
2002
n.a.
2000
2000
n.a.
n.a.
n.a.
n.a.
2004
2002
2006
2001
n.a.
2002
n.a.
2000
2002
2001
2005
n.a.
n.a.
n.a.
2000
2002
n.a.
2000
1999
2000
2000
2003
2002
n.a.

0
0
1
1
1
1
1
1
0
1
0
0
0
1
1
0
1
0
1
1
1
1
0
0
0
1
0
1
1
1
1
1
0
1
1
0
0
0
1
1
0
1
0
1
1
0
0
1

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

1
0
0
1
0
1
1
1
1
1
1
1
1
1
1
1
1
0
0
1
1
1
1
1
0
1
0
0
1
0
1
1
0
1
1
1
1
1
1
0
1
1
0
1
0
1
0
0

1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0

0
1
0
0
1
0
0
1
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0

Kyoto and Carbon Emissions

Table A1. Continued.


Sample

ISO3

Country

Emission growth
(in percent)

Kyoto
(year)

ICC
(year)

Rich

Large

OPEC

EIT

GRC
GTM
GUY
HND
HRV
HUN
IDN
IND
IRL
IRN
ITA
JAM
JOR
JPN
KAZ
KEN
KGZ
KHM
KOR
LAO
LBR
LKA
LTU
LVA
MAR
MDA
MDG
MEX
MKD
MLI
MNG
MOZ
MRT
MUS
MWI
MYS
NAM
NIC
NLD
NOR
NPL
NZL
PAK
PRT
PRY
ROM
RWA
SAU
SDN
SEN

Greece
Guatemala
Guyana
Honduras
Croatia
Hungary
Indonesia
India
Ireland
Iran, Islamic Rep.
Italy
Jamaica
Jordan
Japan
Kazakhstan
Kenya
Kyrgyz Republic
Cambodia
Korea, Rep.
Lao PDR
Liberia
Sri Lanka
Lithuania
Latvia
Morocco
Moldova
Madagascar
Mexico
Macedonia, FYR
Mali
Mongolia
Mozambique
Mauritania
Mauritius
Malawi
Malaysia
Namibia
Nicaragua
Netherlands
Norway
Nepal
New Zealand
Pakistan
Portugal
Paraguay
Romania
Rwanda
Saudi Arabia
Sudan
Senegal

11
32
9
51
15
4
36
28
11
37
5
15
34
1
43
15
9
61
16
50
59
34
0
2
29
11
7
14
5
7
20
56
1
41
14
43
40
19
1
22
11
6
36
2
5
4
5
46
75
37

2002
n.a.
n.a.
n.a.
2007
2002
n.a.
n.a.
2002
n.a.
2002
n.a.
n.a.
2002
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
2003
2002
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
2002
2002
n.a.
2002
n.a.
2002
n.a.
2001
n.a.
n.a.
n.a.
n.a.

2002
n.a.
2004
2002
2001
2001
n.a.
n.a.
2002
n.a.
1999
n.a.
2002
2007
n.a.
2005
n.a.
2002
2002
n.a.
2006
n.a.
2003
2002
n.a.
n.a.
n.a.
2005
2002
2000
2002
n.a.
n.a.
2002
2002
n.a.
2002
n.a.
2001
2000
n.a.
2000
n.a.
2002
2001
2002
n.a.
n.a.
n.a.
1999

1
0
0
0
1
1
0
0
1
1
1
1
0
1
1
0
0
0
1
0
0
0
1
1
0
0
0
1
0
0
0
0
0
1
0
1
0
0
1
1
0
1
0
1
0
1
0
1
0
0

1
1
0
1
0
1
1
1
0
1
1
0
1
1
1
1
1
1
1
1
0
1
0
0
1
0
1
1
0
1
0
1
0
0
1
1
0
1
1
0
1
0
1
1
1
1
1
1
1
1

0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0

0
0
0
0
1
1
0
0
0
0
0
0
0
0
1
0
1
0
0
0
0
0
1
1
0
1
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions

Table A1. Continued.

ISO3

Country

SGP
SLB
SLV
SVK
SVN
SWE
SWZ
SYR
TCD
TGO
THA
TJK
TKM
TTO
TUN
TUR
TZA
UGA
UKR
URY
USA
UZB
VEN
VNM
ZAF
ZMB
ZWE

Singapore
Solomon Islands
El Salvador
Slovak Republic
Slovenia
Sweden
Swaziland
Syrian Arab Rep.
Chad
Togo
Thailand
Tajikistan
Turkmenistan
Trinidad and Tobago
Tunisia
Turkey
Tanzania
Uganda
Ukraine
Uruguay
United States
Uzbekistan
Venezuela, RB
Vietnam
South Africa
Zambia
Zimbabwe

Sample
Emission growth Kyoto ICC
(in percent)
(year) (year) Rich Large OPEC EIT
7
12
11
5
1
3
16
13
110
4
32
20
30
45
23
21
67
62
2
7
4
1
6
77
11
17
36

n.a.
n.a.
n.a.
2002
2002
2002
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
2004
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.

n.a.
n.a.
n.a.
2002
2001
2001
n.a.
n.a.
2006
n.a.
n.a.
2000
n.a.
1999
n.a.
n.a.
2002
2002
n.a.
2002
n.a.
n.a.
2000
n.a.
2000
2002
n.a.

1
0
0
1
1
1
0
0
0
0
1
0
1
1
1
1
0
0
1
1
1
0
1
0
1
0
0

0
0
1
1
0
1
0
1
1
1
1
1
0
0
1
1
1
1
1
0
1
1
1
1
1
1
1

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0

0
0
0
1
1
0
0
0
0
0
0
1
1
0
0
0
0
0
1
0
0
1
0
0
0
0
0

Note: Emission growth is difference between pretreatment and posttreatment (i.e., 1997 to 2000 and
2004 to 2007, respectively) average of log CO2 emissions. Kyoto shows the ratification year of Kyoto
commitment, ICC shows the ratification year of the Rome Statutes of the ICC. EIT is a dummy for
economies in transition.
Due to data availability, the following countries drop out of the sample (compared to the CO2 emissions
data from WDI): Andorra, Antigua and Barbuda, Aruba, Bahamas, Barbados, Belize, Brunei Darussalam, Cape Verde, Central African Republic, Dem. Rep. of the Congo, Dominica, Ecuador, Grenada,
Guinea-Bissau, Haiti, Iceland, Iraq, Israel, Kiribati, Korea Dem. Rep., Kuwait, Lebanon, Liechtenstein,
Luxembourg, Libya, Maldives, Malta, Marshall Islands, Micronesia, Monaco, Montenegro, Myanmar,
Netherlands Antilles, Niger, Nigeria, Oman, Palau, Qatar, Russia, Saint Lucia, Samoa, San Marino, Sao
Tome and Principe, Serbia, Seychelles, Sierra Leone, St. Kitts and Nevis, St. Vincent and the Grenadines,
Suriname, Tonga, Tuvalu, Vanuatu, and Yemen.
Furthermore, the following nonindependent territories are also dropped: American Samoa, Bermuda,
Cayman Islands, Channel Islands, Faroe Islands, French Polynesia, Gibraltar, Greenland, Guam, Hong
Kong, Isle of Man, Kosovo, Macao, Mayotte, New Caledonia, Northern Mariana Islands, Puerto Rico,
Turks and Caicos, Virgin Islands (United States), and West Bank and Gaza.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions

Table A2. Robustness checks: IV estimates on CO2 emissionsfull results.


(A) Placebo
treatment
1999
(A)
Kyoto (0,1)
Kyoto, spatial lag
ln GDP
ln GDP, squared
ln population
ln manufacturinga
ln agriculturea
ln servicea
ln stock of other MEA
Government orientation (0.1, 0.2, 0.3)
Openness, (Exp + Imp)/GDP
WTO (0,1)
Polity (1 to 1)
No. of observations
No. of countries
First-stage diagnostics
Over-ID test (P-value)
Weak-ID test (F-stat)
Second-stage diagnostics
Adj. R2

0.06
(0.04)
0.03
(0.05)
1.37
(1.01)
0.06*
(0.03)
1.34
(1.21)
0.06
(0.07)
0.04
(0.07)
0.11
(0.13)
0.16
(0.15)
0.19
(0.17)
0.11
(0.09)
0.04
(0.08)
0.00
(0.00)
518
133

0.11

(B) Alternative
IV strategy
LIML
(B1)

GMM
(B2)

(C) Kyoto
definition
Stringency
(C)

0.10**
(0.05)
0.02***
(0.00)
0.62
(0.50)
0.00
(0.01)
0.95***
(0.26)
0.12
(0.09)
0.10**
(0.04)
0.10
(0.09)
0.18***
(0.07)
0.15**
(0.06)
0.02
(0.07)
0.00
(0.04)
0.00
(0.00)
1,418
133

0.10**
(0.04)
0.02***
(0.00)
0.58
(0.50)
0.00
(0.01)
0.98***
(0.26)
0.12
(0.09)
0.09**
(0.04)
0.10
(0.09)
0.17**
(0.07)
0.14**
(0.06)
0.02
(0.07)
0.00
(0.04)
0.00
(0.00)
1,418
133

0.05**
(0.02)
0.02***
(0.00)
0.66
(0.51)
0.00
(0.01)
0.99***
(0.25)
0.12
(0.09)
0.09**
(0.04)
0.10
(0.09)
0.17**
(0.07)
0.14**
(0.06)
0.00
(0.07)
0.00
(0.04)
0.00
(0.00)
1,418
133

0.44
19.09

0.44
19.09

0.41
21.94

0.42

0.42

0.42

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions

Table A2. Continued.


(D) Alternative samples
Sample
Kyoto (0,1)
Kyoto, spatial lag
ln GDP
ln GDP, squared
ln population
ln manufacturinga
ln agriculturea
ln servicesa
ln stock of other IEA
Government orientation
Openness, (Exp + Imp)/GDP
WTO (0,1)
Polity (1 to 1)
No. of observations
No. of countries
First-stage diagnostics
Over-ID test (P-value)
Weak-ID test (F-stat)
Second-stage diagnostics
Adj. R2

W/o EIT
(D1)

W/o OPEC
(D2)

Rich only
(D3)

Large only
(D4)

0.11**
(0.05)
0.02***
(0.00)
0.92*
(0.52)
v0.01
(0.01)
0.62**
(0.31)
0.17*
(0.10)
0.12**
(0.05)
0.16
(0.10)
0.14*
(0.08)
0.13**
(0.06)
0.02
(0.08)
0.02
(0.03)
0.00
(0.00)
1,172
110

0.10**
(0.05)
0.02***
(0.00)
0.55
(0.52)
0.00
(0.01)
1.02***
(0.30)
0.12
(0.10)
0.11**
(0.05)
0.07
(0.11)
0.15**
(0.07)
0.14**
(0.06)
0.02
(0.07)
0.00
(0.04)
0.00
(0.00)
1,354
127

0.11***
(0.04)
0.03
(0.02)
0.75
(0.70)
0.00
(0.02)
0.55*
(0.33)
0.32**
(0.15)
0.05
(0.05)
0.03
(0.21)
0.21***
(0.06)
0.06
(0.06)
0.25***
(0.07)
0.02
(0.05)
0.01*
(0.01)
720
67

0.09
(0.06)
0.02***
(0.00)
0.45
(0.58)
0.03*
(0.01)
1.38***
(0.39)
0.03
(0.07)
0.12**
(0.06)
0.14
(0.12)
0.19**
(0.08)
0.13*
(0.07)
0.09
(0.08)
0.01
(0.04)
0.00
(0.00)
1,007
94

0.36
21.37

0.67
18.18

0.10
8.30

0.07
16.28

0.46

0.41

0.43

0.51

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions

Table A2. Continued.


(E) Alternative
treatment window

Kyoto (0,1)
Kyoto, spatial lag
ln GDP
ln GDP, squared
ln population
ln manufacturinga
ln agriculturea
ln servicesa
ln stock of other MEA
Government orientation (0.1, 0.2, 0.3)
Openness, (Exp + Imp)/GDP
WTO (0,1)
Polity (1 to 1)
No. of observations
No. of countries
First-stage diagnostics
Over-ID test (P-value)
Weak-ID test (F-stat)
Second-stage diagnostics
Adj. R2

(F) GHGs

Narrow
(E1)

Broad
(E2)

2000, 2005
(F1)

2000, 2008
(F2)

0.12**
(0.06)
0.04***
(0.01)
0.85
(0.57)
0.01
(0.02)
1.01***
(0.30)
0.15
(0.11)
0.22***
(0.08)
0.24
(0.16)
0.17*
(0.10)
0.19
(0.16)
0.12
(0.13)
0.03
(0.06)
0.01
(0.01)
266
133

0.17**
(0.07)
0.04***
(0.01)
0.63
(0.51)
0.00
(0.01)
1.04***
(0.26)
0.11
(0.08)
0.18**
(0.09)
0.09
(0.14)
0.24**
(0.10)
0.26*
(0.15)
0.06
(0.13)
0.06
(0.06)
0.01
(0.01)
258
129

0.10***
(0.03)
0.00*
(0.00)
0.01
(0.31)
0.02
(0.02)
0.15
(0.30)
0.06
(0.06)
0.00
(0.05)
0.13
(0.11)
0.05
(0.05)
0.00***
(0.00)
0.00
(0.00)
0.07
(0.05)
0.00
(0.00)
218
109

0.27***
(0.10)
0.00**
(0.00)
1.46
(1.06)
0.10*
(0.06)
0.36
(0.46)
0.03
(0.08)
0.07
(0.07)
0.35
(0.33)
0.04
(0.13)
0.00
(0.00)
0.00
(0.00)
0.13
(0.11)
0.00
(0.01)
200
100

0.85
34.06

0.82
37.47

0.48
10.33

0.41
11.59

0.38

0.32

0.14

0.25

Note: Dependent variable is ln CO2 emissions, in Panel F, ln GHG emissions. Long FE estimation. All
regressions use period dummies and constant (not shown). Standard errors in parentheses adjusted for
within-group clustering and heteroskedasticity. *P < 0.1; **P < 0.05; ***P < 0.01. Instruments for Kyoto:
ICC membership and its spatial lag.
a In

percent of GDP.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

1,418
0.48
13.28

0.03
(0.02)
0.00
(0.01)
0.02
(0.01)
0.04***
(0.02)
0.01**
(0.01)
0.65
(0.54)
0.00
(0.01)
1.02***
(0.25)
0.13
(0.10)
0.10**
(0.05)
0.10
(0.10)
0.18**
(0.07)
0.15**
(0.06)
0.01
(0.08)
0.00
(0.04)
0.00
(0.00)

CO2 emission

1,180
0.29
6.11

0.80
(0.39)
0.42**
(0.21)
0.26
(0.20)
0.57
(0.36)
0.59***
(0.18)
1.02
(9.94)
0.14
(0.28)
15.25*
(7.70)
1.49
(1.21)
1.74
(1.28)
0.82
(2.54)
1.45
(1.14)
0.04
(1.72)
0.12
(1.30)
0.42
(0.72)
0.11*
(0.06)

**

Renewables

1,180
0.17
5.74

0.56
(0.53)
0.42
(0.39)
0.28
(0.46)
0.94**
(0.45)
0.53**
(0.23)
1.25
(10.95)
0.24
(0.30)
15.42**
(7.66)
0.14
(1.41)
2.14
(1.36)
1.73
(2.74)
2.03
(1.44)
3.60*
(1.94)
0.14
(1.33)
0.21
(1.03)
0.07
(0.07)

Fossils

Share of

1,180
0.06
1.86

0.03
(0.67)
0.80
(0.72)
0.28
(0.60)
0.93*
(0.53)
0.21
(0.19)
7.65
(10.58)
0.35
(0.29)
6.17
(6.19)
2.53*
(1.31)
0.64
(1.13)
1.12
(1.17)
1.44
(1.35)
0.18
(2.53)
1.91
(1.31)
0.13
(0.91)
0.17
(0.12)

Coal

1,180
0.21
2.73

0.50
(0.18)
0.07
(0.24)
0.57**
(0.24)
1.23***
(0.31)
0.01
(0.06)
0.12
(3.35)
0.04
(0.10)
2.67
(1.88)
0.14
(0.49)
0.01
(0.33)
0.08
(0.51)
0.62
(0.47)
0.10
(1.22)
0.25
(0.35)
0.24
(0.19)
0.06
(0.06)

***

Clean

610
0.79
57.50

0.01
(0.03)
0.05
(0.10)
0.14
(0.10)
0.03
(0.03)
0.01*
(0.01)
1.22**
(0.51)
0.03**
(0.01)
0.19
(0.30)
0.03
(0.05)
0.05
(0.06)
0.21**
(0.09)
0.08
(0.07)
0.15
(0.13)
0.01
(0.07)
0.02
(0.04)
0.00
(0.00)

Diesel

610
0.70
51.12

0.05
(0.03)
0.05
(0.07)
0.14**
(0.07)
0.05
(0.03)
0.01
(0.01)
0.55
(0.61)
0.01
(0.02)
0.35
(0.36)
0.02
(0.07)
0.05
(0.09)
0.27**
(0.11)
0.11
(0.15)
0.19
(0.15)
0.08
(0.10)
0.02
(0.05)
0.00
(0.00)

Gasoline

Price of

1,180
0.39
7.48

0.03
(0.01)
0.00
(0.01)
0.02**
(0.01)
0.05***
(0.01)
0.00
(0.00)
0.57*
(0.32)
0.02**
(0.01)
0.43**
(0.18)
0.04
(0.04)
0.06*
(0.04)
0.12*
(0.07)
0.02
(0.05)
0.01
(0.05)
0.06*
(0.03)
0.02
(0.02)
0.00
(0.00)
*

Energy

1,169
0.52
14.25

0.03
(0.02)
0.01
(0.01)
0.01
(0.01)
0.03*
(0.02)
0.00
(0.01)
0.04
(1.05)
0.01
(0.03)
0.51
(0.35)
0.20*
(0.12)
0.01
(0.05)
0.08
(0.08)
0.16
(0.12)
0.10
(0.08)
0.12
(0.10)
0.08*
(0.05)
0.00
(0.00)

Electricity

ln of per capita use of

Note: FE estimations with time lags. Standard errors in parentheses adjusted for within-group clustering and heteroskedasticity. *P < 0.1; **P < 0.05; ***P < 0.01. All regressions use
year dummies and constant (not shown).

Observations
Adj. R2
F-statistic

Polity (1 to 1)

WTO (0,1)

Openness, (Exp + Imp)/GDP

Government orientation

ln stock of other MEA

ln services

ln agriculture

ln manufacturing

ln population

ln GDP, squared

ln GDP

Kyoto, spatial lag

Kyoto3 (0,1)

Kyoto2 (0,1)

Kyoto1 (0,1)

Kyoto (0,1)

Dep. var.

Table A3. Kyoto effects with lag structurefull results.

Kyoto and Carbon Emissions

No. observations (countries)


Over-ID test (P-value)
Weak-ID test (F-stat)
Adj. R2

Polity (1 to 1)

WTO (0,1)

Openness

Government orientation

ln MEA stock

ln services

ln agriculture

ln manufacturing

ln population

ln GDP, squared

ln GDP

Kyoto, spatial lag

Kyoto (0,1)

Method

Dep. var.

Panel

2.41***
(0.93)
0.62***
(0.15)
0.02
(9.24)
0.17
(0.26)
13.31*
(7.80)
1.35
(1.14)
1.76
(1.17)
0.74
(2.40)
1.40
(1.06)
0.01
(1.61)
0.07
(1.17)
0.36
(0.67)
0.11*
(0.06)
1,180 (110)
0.68
18.80
0.21

1.38**
(0.56)
0.58***
(0.18)
1.16
(9.99)
0.15
(0.28)
16.89**
(7.45)
1.34
(1.25)
1.58
(1.28)
0.77
(2.53)
1.47
(1.16)
0.01
(1.71)
0.18
(1.29)
0.53
(0.71)
0.11*
(0.06)
1,180 (110)
0.29

FE-IV

FE-OLS

Renewables

2.46**
(1.16)
0.61***
(0.16)
3.15
(10.32)
0.29
(0.28)
10.58
(8.05)
0.26
(1.30)
2.31*
(1.24)
1.64
(2.62)
1.94
(1.37)
3.59**
(1.82)
0.30
(1.25)
0.01
(0.99)
0.07
(0.07)
1,180 (110)
0.47
18.80
0.06

0.67
(0.63)
0.52**
(0.22)
1.18
(11.02)
0.25
(0.30)
16.78**
(7.46)
0.28
(1.45)
2.00
(1.35)
1.69
(2.74)
2.05
(1.45)
3.60*
(1.91)
0.14
(1.33)
0.31
(1.02)
0.07
(0.07)
1,180 (110)
0.17

FE-IV

FE-OLS

Fossil fuel

(A) Shares in energy use

Table A4. Through which channels does Kyoto operate? Full results.

0.06

0.12
(0.93)
0.22
(0.19)
7.62
(10.58)
0.36
(0.29)
7.14
(5.58)
2.41*
(1.30)
0.76
(1.15)
1.14
(1.16)
1.46
(1.35)
0.17
(2.57)
1.94
(1.31)
0.06
(0.94)
0.17
(0.12)
1,180 (110)

FE-OLS

FE-IV
1.43
(1.76)
0.14
(0.11)
9.33
(10.36)
0.40
(0.28)
1.77
(7.93)
2.43**
(1.21)
0.49
(1.12)
1.19
(1.13)
1.36
(1.29)
0.16
(2.43)
2.32*
(1.40)
0.33
(1.00)
0.17
(0.11)
1,180 (110)
0.60
18.80
0.05

Coal

0.17

1.07***
(0.27)
0.03
(0.06)
0.30
(3.40)
0.05
(0.10)
0.12
(1.91)
0.39
(0.51)
0.26
(0.37)
0.15
(0.52)
0.57
(0.47)
0.14
(1.33)
0.19
(0.39)
0.42**
(0.20)
0.06
(0.06)
1,180 (110)

FE-OLS

1.66***
(0.56)
0.00
(0.03)
0.35
(3.37)
0.06
(0.10)
2.18
(2.59)
0.39
(0.46)
0.16
(0.36)
0.17
(0.49)
0.61
(0.43)
0.13
(1.24)
0.34
(0.42)
0.32
(0.22)
0.06
(0.05)
1,180 (110)
0.63
18.80
0.08

FE-IV

Alternative energy

(B) Shares in electricity production

Kyoto and Carbon Emissions

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

0.68

0.13***
(0.04)
0.01
(0.01)
0.55
(0.60)
0.01
(0.02)
0.12
(0.33)
0.02
(0.08)
0.04
(0.09)
0.28**
(0.11)
0.11
(0.15)
0.21
(0.16)
0.10
(0.10)
0.00
(0.04)
0.00
(0.00)
608 (127)

FE-OLS
0.25***
(0.05)
0.00
(0.01)
0.62
(0.55)
0.02
(0.01)
0.56
(0.35)
0.01
(0.06)
0.06
(0.08)
0.28***
(0.09)
0.10
(0.13)
0.20
(0.14)
0.08
(0.09)
0.02
(0.04)
0.00
(0.00)
608 (127)
0.32
20.30
0.58

FE-IV

Gasoline

0.05*
(0.03)
0.00
(0.00)
0.57*
(0.30)
0.02**
(0.01)
0.34*
(0.19)
0.03
(0.04)
0.05
(0.03)
0.13*
(0.07)
0.02
(0.05)
0.01
(0.05)
0.06*
(0.03)
0.03
(0.02)
0.00
(0.00)
1,180 (110)
0.36
18.80
0.31

0.05**
(0.02)
0.00
(0.00)
0.56*
(0.32)
0.02**
(0.01)
0.33*
(0.18)
0.03
(0.04)
0.05
(0.04)
0.13*
(0.07)
0.02
(0.05)
0.01
(0.05)
0.06*
(0.03)
0.03
(0.02)
0.00
(0.00)
1,180 (110)
0.38

FE-IV

FE-OLS

Energy

0.52

0.04
(0.03)
0.00
(0.01)
0.04
(1.05)
0.01
(0.03)
0.58*
(0.32)
0.20
(0.12)
0.01
(0.04)
0.07
(0.08)
0.16
(0.12)
0.10
(0.08)
0.12
(0.10)
0.08*
(0.05)
0.00
(0.00)
1,169 (109)

FE-OLS

0.08**
(0.04)
0.00
(0.01)
0.00
(0.96)
0.01
(0.02)
0.44
(0.29)
0.20*
(0.12)
0.01
(0.04)
0.07
(0.08)
0.16
(0.11)
0.10
(0.08)
0.11
(0.09)
0.08*
(0.04)
0.00
(0.00)
1,169 (109)
0.21
18.84
0.47

FE-IV

Electricity

(D) Log per capita use of

Note: FE estimation. All regressions use year dummies and constant (not shown). Standard errors in parentheses adjusted for within-group clustering and
heteroskedasticity. *P < 0.1; **P < 0.05; ***P < 0.01. Instruments for Kyoto: ICC membership and its spatial lag.

No. of observations (countries)


Over-ID test (P-value)
Weak-ID test (F-stat)
Adj. R2

Polity (1 to 1)

WTO (0,1)

Openness

Gov. orientation

ln MEA stock

ln services

ln agriculture

ln manufacturing

ln population

ln GDP, squared

ln GDP

0.77

0.22***
(0.05)
0.01
(0.00)
1.30***
(0.46)
0.04***
(0.01)
0.01
(0.28)
0.04
(0.05)
0.05
(0.05)
0.22***
(0.08)
0.06
(0.06)
0.17
(0.11)
0.01
(0.06)
0.02
(0.03)
0.00
(0.00)
608 (127)
0.61
20.30
0.69

0.10**
(0.04)
0.01*
(0.01)
1.23**
(0.52)
0.04**
(0.01)
0.45
(0.28)
0.03
(0.06)
0.03
(0.06)
0.22**
(0.09)
0.07
(0.07)
0.17
(0.13)
0.03
(0.07)
0.01
(0.04)
0.00
(0.00)
608 (127)

Kyoto (0,1)

Kyoto, spatial lag

FE-IV

FE-OLS

Diesel fuel

(C) Pump prices (USD/L)

Method

Dep. var.

Panel

Table A4. Continued.

Kyoto and Carbon Emissions

Kyoto and Carbon Emissions

Table A5. Robustness checks: channels. IV estimatesfull results.


(A) Shares in energy use
Dep. var.

Renewables

Fossil fuel

Method/sample

Long FE
(A1)

W/o EIT
(A2)

Rich only
(A3)

Long FE
(A4)

W/o EIT
(A5)

Rich only
(A6)

Kyoto (0,1)

3.43***
(1.14)
0.81***
(0.21)
4.40
(9.35)
0.29
(0.26)
7.24
(6.32)
3.06
(2.15)
2.09
(1.79)
0.44
(4.47)
1.71
(1.55)
1.11
(4.85)
2.48
(2.45)
0.40
(1.09)
0.16
(0.12)
220
110
0.32
36.49
0.32

3.44***
(1.29)
0.66***
(0.13)
6.24
(10.95)
0.06
(0.29)
4.68
(12.33)
1.43
(1.69)
2.07
(1.44)
0.09
(3.19)
0.08
(1.76)
0.42
(2.06)
0.89
(1.42)
0.32
(1.01)
0.12*
(0.06)
923
86
0.53
15.81
0.24

2.05**
(0.85)
0.07
(0.22)
5.33
(16.52)
0.24
(0.43)
0.82
(5.22)
1.31
(1.86)
0.58
(0.78)
3.37
(2.45)
1.20
(1.52)
0.71
(1.31)
0.48
(0.83)
0.32
(0.45)
0.02
(0.04)
701
65
0.72
8.39
0.05

3.62***
(1.37)
0.81***
(0.22)
8.89
(9.74)
0.41
(0.28)
2.93
(6.65)
4.25
(3.13)
3.55*
(1.90)
0.44
(4.67)
1.70
(2.04)
2.19
(5.19)
4.66*
(2.57)
0.19
(1.34)
0.13
(0.14)
220
110
0.18
36.49
0.58

2.80**
(1.29)
0.67***
(0.12)
0.74
(10.47)
0.26
(0.27)
2.30
(12.13)
0.00
(1.71)
2.49*
(1.44)
0.26
(3.30)
0.71
(1.81)
4.27*
(2.20)
1.09
(1.40)
0.44
(0.96)
0.09
(0.08)
923
86
0.50
15.81
0.12

0.58
(1.13)
0.79**
(0.40)
0.06
(18.08)
0.16
(0.45)
1.72
(6.18)
1.83
(1.96)
0.08
(0.89)
2.30
(2.75)
2.73
(2.15)
4.45***
(1.70)
0.17
(1.04)
0.15
(0.54)
0.03
(0.06)
701
65
0.58
8.39
0.02

Kyoto, spatial lag


ln GDP
ln GDP, squared
ln population
ln manufacturing share
ln agriculture share
ln services share
ln IEA stock
Government orientation
Openness
WTO (0,1)
Polity (1 to 1)
No. of observations
No. of countries
Over-ID test (P-value)
Weak-ID test (F-stat)
Adj. R2

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions

Table A5. Continued.


(B) Shares in electricity production
Dep. var.

Alternative energy

Coal

Method/sample

Long FE
(B1)

W/o EIT
(B2)

Rich only
(B3)

Long FE
(B4)

W/o EIT
(B5)

Rich only
(B6)

Kyoto (0,1)

3.06***
(0.58)
0.05
(0.06)
3.21
(4.52)
0.12
(0.13)
8.35***
(2.18)
0.28
(0.82)
0.20
(0.72)
0.16
(1.08)
1.19*
(0.69)
0.91
(2.54)
1.19
(0.87)
0.19
(0.43)
0.09
(0.09)
220
110
0.60
36.49
0.47

2.06***
(0.56)
0.00
(0.03)
1.95
(3.72)
0.10
(0.11)
0.84
(3.35)
0.78
(0.74)
0.07
(0.40)
0.87
(0.66)
0.35
(0.61)
0.20
(1.56)
0.52
(0.52)
0.33*
(0.19)
0.09
(0.06)
923
86
0.55
15.81
0.09

1.69***
(0.54)
0.12
(0.16)
2.09
(5.76)
0.00
(0.14)
1.16
(3.56)
0.63
(0.52)
0.34
(0.47)
0.16
(0.94)
0.95
(0.76)
0.26
(1.60)
0.03
(0.66)
0.00
(0.33)
0.03
(0.02)
701
65
0.30
8.39
0.21

0.96
(1.41)
0.14
(0.11)
11.34
(12.82)
0.48
(0.34)
1.89
(4.27)
6.66**
(2.59)
0.36
(2.14)
2.98
(2.13)
1.07
(2.11)
2.62
(5.91)
7.08**
(3.00)
0.92
(1.30)
0.27
(0.17)
220
110
0.62
36.49
0.74

1.71
(2.29)
0.10
(0.07)
6.34
(12.77)
0.33
(0.32)
0.66
(8.96)
3.51*
(1.84)
0.82
(1.27)
1.76
(1.31)
0.52
(1.57)
1.65
(2.81)
2.31
(1.57)
0.54
(0.83)
0.08
(0.09)
923
86
0.47
15.81
0.06

0.74
(1.30)
1.32*
(0.69)
7.10
(13.14)
0.21
(0.34)
2.29
(8.60)
2.06
(1.29)
1.51
(1.38)
0.07
(1.95)
0.88
(2.42)
1.44
(3.35)
2.28
(1.62)
0.88
(1.14)
0.37
(0.24)
701
65
0.92
8.39
0.05

Kyoto, spatial lag


ln GDP
ln GDP, squared
ln population
ln manufacture share
ln agriculture share
ln services share
ln IEA stock
Government orientation
Openness
WTO (0,1)
Polity (1 to 1)
No. of observations
No. of countries
Over-ID test (P-value)
Weak-ID test (F-stat)
Adj. R2

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions

Table A5. Continued.


(C) Pump prices
Dep. var.

Diesel fuel

Gasoline

Method/sample

Long FE
(C1)

W/O EIT
(C2)

Rich only
(C3)

Long FE
(C4)

W/o EIT
(C5)

Rich only
(C6)

Kyoto (0,1)

0.49***
(0.06)
0.01
(0.01)
0.98*
(0.54)
0.02
(0.01)
1.17***
(0.23)
0.00
(0.10)
0.06
(0.08)
0.18
(0.12)
0.11
(0.10)
0.25
(0.23)
0.13
(0.09)
0.14*
(0.07)
0.01**
(0.01)
252
126
0.04
39.06
0.47

0.12**
(0.05)
0.01
(0.00)
0.98*
(0.53)
0.03*
(0.01)
0.11
(0.39)
0.01
(0.05)
0.03
(0.05)
0.27***
(0.09)
0.08
(0.09)
0.20
(0.14)
0.04
(0.07)
0.05
(0.04)
0.00
(0.00)
498
104
0.67
18.86
0.67

0.31***
(0.08)
0.07*
(0.04)
0.67
(0.71)
0.02
(0.02)
0.07
(0.45)
0.00
(0.08)
0.14**
(0.06)
0.48***
(0.16)
0.24*
(0.13)
0.21*
(0.13)
0.06
(0.08)
0.05
(0.04)
0.01*
(0.01)
316
65
0.39
8.83
0.67

0.46***
(0.06)
0.00
(0.01)
0.55
(0.60)
0.01
(0.02)
1.34***
(0.26)
0.07
(0.10)
0.03
(0.10)
0.20
(0.14)
0.05
(0.13)
0.24
(0.21)
0.05
(0.13)
0.14*
(0.08)
0.01
(0.01)
252
126
0.47
39.06
0.40

0.17***
(0.06)
0.00
(0.01)
0.27
(0.67)
0.01
(0.02)
0.39
(0.42)
0.06
(0.09)
0.04
(0.09)
0.31***
(0.10)
0.20
(0.17)
0.26
(0.17)
0.11
(0.10)
0.03
(0.06)
0.00
(0.00)
498
104
0.34
18.86
0.56

0.32***
(0.09)
0.06*
(0.03)
0.28
(0.78)
0.01
(0.02)
0.44
(0.59)
0.02
(0.09)
0.05
(0.10)
0.51***
(0.18)
0.12
(0.12)
0.26*
(0.14)
0.11
(0.08)
0.06
(0.05)
0.01
(0.01)
316
65
0.04
8.83
0.60

Kyoto, spatial lag


ln GDP
ln GDP, squared
ln population
ln manufacturing share
ln agriculture share
ln services share
ln IEA stock
Government orientation
Openness
WTO (0,1)
Polity (1 to 1)
No. of observations
No. of countries
Over-ID test (P-value)
Weak-ID test (F-stat)
Adj. R2

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Kyoto and Carbon Emissions

Table A5. Continued.


(D) Per capita use of
Dep. var.

Energy

Electricity

Method/sample

Long FE
(D1)

W/o EIT
(D2)

Rich only
(D3)

Long FE
(D4)

W/o EIT
(D5)

Rich only
(D6)

Kyoto (0,1)

0.06
(0.04)
0.00
(0.00)
0.73***
(0.27)
0.03***
(0.01)
0.25
(0.20)
0.03
(0.07)
0.10*
(0.05)
0.24**
(0.11)
0.03
(0.07)
0.11
(0.14)
0.13*
(0.07)
0.08**
(0.03)
0.00
(0.00)
220
110
0.67
36.49
0.02

0.07*
(0.04)
0.00
(0.00)
0.61*
(0.33)
0.02**
(0.01)
0.54**
(0.27)
0.10**
(0.04)
0.08**
(0.03)
0.19***
(0.06)
0.04
(0.07)
0.06
(0.06)
0.07*
(0.04)
0.02
(0.02)
0.00
(0.00)
923
86
0.31
15.81
0.36

0.06*
(0.03)
0.01
(0.01)
0.47
(0.38)
0.02**
(0.01)
0.38
(0.24)
0.05
(0.05)
0.01
(0.04)
0.12
(0.12)
0.04
(0.04)
0.06
(0.04)
0.00
(0.04)
0.01
(0.03)
0.01
(0.00)
701
65
0.42
8.39
0.43

0.02
(0.06)
0.00
(0.01)
0.21
(0.94)
0.02
(0.02)
0.94***
(0.25)
0.23
(0.15)
0.10
(0.08)
0.07
(0.15)
0.16
(0.14)
0.20
(0.22)
0.23
(0.20)
0.11
(0.08)
0.01
(0.01)
218
109
0.92
36.30
0.15

0.13***
(0.04)
0.00
(0.01)
1.21
(1.24)
0.02
(0.03)
0.29
(0.28)
0.37**
(0.18)
0.08
(0.05)
0.08
(0.10)
0.16
(0.19)
0.08
(0.09)
0.11
(0.11)
0.11
(0.08)
0.00
(0.00)
912
85
0.03
16.23
0.53

0.06
(0.04)
0.00
(0.01)
1.02
(0.64)
0.04**
(0.02)
0.26
(0.24)
0.00
(0.05)
0.02
(0.04)
0.11
(0.10)
0.04
(0.05)
0.06
(0.06)
0.05
(0.05)
0.03
(0.03)
0.00
(0.01)
701
65
0.74
8.39
0.65

Kyoto, spatial lag


ln GDP
ln GDP, squared
ln population
ln manufacturing share
ln agriculture share
ln services share
ln IEA stock
Government orientation
Openness
WTO (0,1)
Polity (1 to 1)
No. of observations
No. of countries
Over-ID test (P-value)
Weak-ID test (F-stat)
Adj. R2

Note: FE IV estimates. Standard errors in parentheses adjusted for within-group clustering and heteroskedasticity. All regressions use year dummies and constant (not shown). *P < 0.1; **P < 0.05; ***P
< 0.01. Instruments for Kyoto: ICC membership and its spatial lag.

Journal of Policy Analysis and Management DOI: 10.1002/pam


Published on behalf of the Association for Public Policy Analysis and Management

Potrebbero piacerti anche